Economics of Education Midterm 2

Total Expenditure per Student (2010-2011)

$12,608

Israel Class Study

-2 Teachers employed if there were more than 40 students in a grade level

School Choice Policy Discuss the economic arguments for and against the idea that school choice policies will lead to more efficient education outcome

-For: different types of students/parents benefit from different allocation of resources, so more choices increases competition, and would be better for all students -Against: Sometimes, there will be a school that clearly allocates resources more efficiently, so there will be a higher demand, leaving the other option with a demand of close to 0. This results in an inefficient imbalance of students anding each school -education isn't a competitive market so moving public funding resources and students away from public schools will hurt the public schooling system

School choice leads to competition and then gains in efficiency. What do we need for this to occur?

-Informed parents -No barriers to entry or exit -A choice

Class Size: What are the findings from Hoxby (2000)? How can we reconcile her findings with those from Project STAR? What is the implication of the difference between these studies for using the Project STAR results to support lower class size mandates?

-Leverages class size variation due to population variation across 650 Connecticut elementary schools, combined with max and min class size rules that trigger large class size changes from small variation in school size -Finds no evidence that smaller class size leads to higher test scores cast doubt on true impact -STAR program schools had to be big enough for a small class and a large class, meaning no new teachers needed to be hired. HOXBY's they likely needed to hire new teachers who are usually less experienced, and have low performance.

Proponents of charter school and open enrollment policies often argue that choice policies will be beneficial to students who remain in their assigned public school. Opponents of these policies argue it will harm public schools. What is the basis for both of these arguments? What do the empirical studies on school choice have to say about which side is right?

-Proponents of choice policy: increasing school choice will inject more competitive forces into public education benefitting all students - Critics of choice policy: education is not a comp. market, so moving public funding resources and students away from traditional public schools will harm the public education system more broadly These studies support the existence of a + link b/w competition and public school productivity -Research impact on competition is mixed/+

Class Size Summary

-Students from smaller classes achieve better test scores -Teaching Assistants do not appear to have a beneficial effect -Both RCTs and Natural Experiments provide similar results -Reducing class size by 7 increases Kindergarten test scores -Although when class size reduction policies have been introduced find negligible results Angrist (2004) "under relatively ideal circumstances (in particular, given long enough lead time and funding adequate to reduce class size without compromising other educational inputs), class size reductions can increase test scores. This is an important result and a performance standard that many and perhaps most programs, however well-intentioned, do not live up to."

Summarize the responses to the Hanushek Critique. (5)

1) Targeted spending: A major target of the school finance reform movement was to equalize spending inequities across districts, and if spending was equalized, we would see gains predominantly for students from lower-SES backgrounds, not the average ---Closed the achievement gap of test scores between whites and minorities (blacks/hispanics) suggesting that changes in school financing had a result 2) Compositional Differences: change in types of students in US schools/tested (fewer students dropping out, and keeping them in school likely to bring down average test score) ---Large increases in proportion of hispanic immigrants, number of students with disabilities, and children from single-headed households are consistent with this problem of compositional difference, as they are often harder to educate and therefore, more expensive to educate. ---Failing to account for the changes in student population and the expansion in the range of services provided could be seen as a problem of omitted variables. *NOTE: Test increases amongst younger students, but not seen in 17-year olds. 3) Test Scores Poor Measure of Ability: A single test may be unable to capture all the educational gains - breadth and depth of skill acquisition through education -Researchers focus on Long-Run outcomes --Card and Krueger - changes in school resources in 20s-30s and earnings in 1980s = find large impact of resources on later earnings --Arronson and Mazumder - Spread of Rosenwald schools in 1913-31 (schools for black children in south, of much higher quality and more financial resources that other schools for blacks) ---These schools had large effect on educational attainment of students, as well as on literacy and northern migration ------These are closely tied to earning, so likely their earnings were higher 4) Limitations in Literature Review: literature reviews are not useful for determining causal relationships because they contain observational studies, random control tests, and natural experiments Review is not formal meta analysis, which would: 1.Put coefficients into common metric 2. Studies with larger samples should get greater weight 3. Repeated analysis of same sample should not be counted multiple times 4. All studies are not created equal 5) Range of Results

8. The archipelago nation of Tuvalu has enlisted you to study their school choice policies. Because the nation consists of a series of small islands, typically there was little school choice. However, the government recently implemented a charter school policy that led to a large increase in charter school prevalence. Assume you have longitudinal data at the student level over several years on student math test scores, demographic characteristics, and what school each student enrolls in each year. 1. What would be the problem with simply comparing the math test scores of students attending a traditional public school with the outcomes of students who attend a charter school, even controlling for student demographic characteristics? Would this comparison yield the causal effect of attending a charter school on math test scores? 2. Some charter schools are over-subscribed, and by law they are required to admit people by lottery. How would you use these lottery data to overcome the selection problems you discussed in part (a)? Would this method tell you how an average charter school in Tuvalu affects student math test scores? 3. Using a variety of methods, you find that charter school enrollment leads to large increases in student math tests. The government of Tuvalu uses this information to argue that they should expand charter school access to more students. Does this policy conclusion necessarily follow from your results? Explain.

1. -Charter schools are unable to control the set of students who enroll so they are unable to focus on teaching specific types of students -Charter schools tend to be in low-income areas with high demand for education (more blacks and hispanics, inner-city areas, bad experiences in public schools), which can lead to a self-selection problem. There this would not lead to a causal effect because these other variables could be playing a role 2. a. You can look at the gains in the test scores, not just the test scores alone b. You could use the lottery as a way to control and estimate causal effect. The lottery creates random differences in school options amongst those who apply c. It could, but you have to control for things like socioeconomic factors that drive students away from publicly assigned schools and into alternative charter schools. The method could just measure the effect of charter schools on a certain type of student and not be representative of student scores as a whole no external validity d. The over subscribed schools may be the best charter schools estimates lack external validity 3. -Yeah, it does follow it but it's still not a concrete correlation -Also, more schools means more competition, which could cut back on segregation issues in charter schools

Rhonda is the manager of a plant that produces cars, while Sofia is the principal of a local elementary school. 1. What is Rhonda's main objective as the manager of the car plant? 2. Assuming Rhonda has full knowledge of the production function for cars, what will she do to meet this objective? 3. How is Sofia's objective as principal similar or different from Rhonda's?Explain the differences in the way knowledge relative to cars is produced that make it more challenging for Sofia to meet her objectives.

1. Choose inputs to maximize profits (aluminum, steel, plastics, wheels) improve profits either by: -producing the same number of cars at a lower cost -produce higher quality cars with the same inputs 2. Rhonda must choose inputs such as capital and labor to maximize profits with the given production function 3. Similarities: They both want to combine inputs to produce an optimal output Differences: Sofia doesn't know what all her relevant inputs are/each inputs degree of relevance -importance of teacher quality, computers, class size, textbooks, peer quality, etc. Measuring output directly Quantifying outputs -With education is it difficult to measure outputs -Rhonda want to max her profit -Sofia (assuming the school is not-for-profit) wants to create cost minimization

Reason for Small Classes

1. Compensatory resource allocation: (provides more resources to students who have low levels of education resources due to out-of-school reasons) Academically challenged students put into specialist classes High resources for students from low income backgrounds 2. Complementary resource allocation (provides more resources to students who have high levels of education resources due to out-of-school reasons) Academically brilliant students put into specialist classes High resources for students from high income backgrounds 3. Other resource allocation (provides more resources to students regardless of education resources due to out-of-school reasons) Lots of births one year, few another year Retirement, death of teachers *Any compensatory or complementary reasons will bias the results*

Consider the problem of how teachers can allocate their time between group instruction (G) and individual instruction (I). The number of students in the class (N) is 30. The class time is 300 minutes per week. Assume that each student gets the same amount of time of individual instruction (I). 1. What is the budget constraint faced by the teachers? 2. What happens to the budget constraint if we increase total instruction time by 50 minutes? 3. What happens to the budget constraint if we decrease class size to 20 students?

1. I + G = 300 2. It shifts the budget constraint to the right 3. The slope becomes flatter *There would be no effect on the budget constraint if using total individual instruction time on the budget constraint, because G+I=300 no matter what. If using individual instruction time per person on budget constraint, then the slope will become flatter, since individual instruction time per person(i)=I/N.

School Choice Policies: Inter-Intra-District Open Enrollment

1. Inter-intra-district open enrollment: No additional schools, allow students choice from amongst existing schools Intra District Competition - Any school within the district Inter District Competition - Any school in the state Common in other countries (UK Sweden) Schools cannot select students, but ability to enrol depends on their being space available Barriers Remain high for non-traditional schools Choice Set Do not necessarily provide transport Low natural number of schools

Consider an education production function with two inputs: class size and books. 1. Draw the budget constraint associated with these two inputs, with books on the x-axis. What determines the slope of the budget constraint? 2. Draw the isoquant that shows the efficient allocation. What is the slope of the isoquant? What property of production functions gives it its shape 3. Why can't the school achieve a point above the budget constraint 4. Show what happens the efficient allocation if the school district get more overall resources. What if the price of books rises?

1. Negative Ratio of prices between two inputs (-Pb/Pc) 2. The ratio of the MP between the 2 goods (-MPb/MPc). Shape reflects diminishing marginal productivity of each input 3. Because anything above the line is unattainable given the level of total resources available to them 4. More resources would push the production function out and to the right Would decrease the number of books that could be attained(become steeper)

Imagine you have been made the superintendent of a school district. As part of your job, you need to allocate resources among four different inputs: teachers, books, pencils, and smart boards. Each one of these inputs has a price associated with it, denoted Pt, Pb, Pp and Ps, respectively. 1. Explain what the marginal product of books is. 2. Holding the number of teachers, pencils and smart boards constant, will the marginal product of books increase or decrease as you add more books? Explain. 3. Write the formula that shows how marginal product and prices will be related amongst all four inputs when allocation is efficient. Describe in words what this efficiency condition means. 4. Is providing the best teachers to the weakest student an example of a complementary or a compensatory policy?

1. Shows the relationship b/w the educational output altering the amount of inputs by 1 unit, holding all other inputs fixed- 2. -Decrease because diminishing marginal product (increasing one input while holding the others constant will increase output by successively smaller amounts) -- the first book will yield the greatest return 3. a. MPt/pt = MPb/pb = MPp/pp = MPs/ps b. Production Possibilities Frontier 4. -Compensatory

Consider the various outputs of the education process 1. List three outputs that are easy to measure 2. List three outputs that are difficult to measure 3. How does the existence of these difficult-to-measure outputs complicate our understanding of the education production function?

1. Test scores, GPA, attendance rate 2. Knowledge, skills, character 3. Since they are difficult to measure, it's hard to tell what inputs really affect these outputs. It is often ambiguous and even if a conclusion is formed, it could be bias.

Compositional Difference Calculation

100 residents of the country 80 from advantaged group 20 from disadvantaged group Advantaged literacy rate 91% Disadvantaged literacy rate 75% Therefore total of 88 residents are literate in English. (73+15)

Iso-quant

A line through which the same quantity of output is produced while changing the quantities of two or more inputs

Value Added Models (VAM) - Validity 1. Teachers are unaffected by their working environment 2. Pupil assignments to teachers are random once the prior test score is taken into account 3. Test scales are invariant 4. Teachers are equally effective with all pupils

ALL ARE UNTRUE: 1. Basic VAM attributes all the gains or lack of to the teacher However other workplace factors may have impacts on student growth Solutions: Account for measures of school quality e.g. control for infrastructure Make comparisons within schools e.g. use of school fixed effects effectively compares the growth in test scores across teachers with pupils of similar backgrounds 2. VAMs use previous test scores as a base upon which the current teacher must build and that the difficulty of teaching students does not vary non-randomly given this test score If teachers are matched with a non-random set of pupils, then value added measures would be biased Solutions Account for student characteristics e.g. prior behaviour 3. VAMs assume that the gains in human capital from improving the test score by one point are the same at all points across the possible score distribution. The gains of improving from 5 to 10 are equivalent to moving from 65 to 70 and 90 to 95. The psychometricians who scale tests do not make these claims Solutions Allow for the growth to be different for each test score e.g. control for each test score 4. VAMs assume that teachers are equally effective with all types of pupils Sharing race, gender and ethnicity all decrease the chances of the teacher finding a pupil disruptive, and increase test scores (Dee 2004/2005) Increases in the likelihood of female students graduating in a STEM subject when taught by a female professor (Carrell et al. 2010) "High performing students benefit from high cognitive teachers, being matched to such a teacher can even be detrimental to their lower performing peers" Grönqvist and Vlachos (2008) If all pupils were sorted such that each was allocated to a teacher that was highly effective with them, all teachers would have higher value-added estimates compared to if pupils were randomly allocated to teachers McCaffrey (2009) found that these match effects account for less than 10% of variation in teacher value added estimates.

School Choice Policies - Private Schools

Allow for religious education 10% of K-12 students enrolled in private schools, Cities differ in the number of private schools for historical reasons Chicago had immigrants from Catholic countries (Ireland/Poland) Barriers Accreditation standards Private/religious donors mandate certain restrictions Choice set Increase choice for parental taste preferences f(.) Increase choice for parental resource preferences (s)

School Choice Policy Traditionally, Hobbits have relied on private provision for their schooling. As the newly-elected mayor of the Shire, Samwise Gamgee has decided to launch a public schooling option. The typical Hobbit family has the following budget constraint: I=S+X, where I is total family income, S is schooling expenditures and X is expenditures on all other goods. 1. Draw the Hobbit family's budget constraint prior to the opening of the public school. 2. Hobbits are notoriously tax-averse, and as a result the level of schooling services offered by the public school is small. Draw the new budget constraint the Hobbit family will face 3. Show an example where the level of education services consumed by the Hobbit family declines due to the public school Samwise is concerned that the amount of schooling has declined. He therefore closes the public school and offers all students a $1,000 voucher that can only be used on education. Draw the new budget constraint. Will the voucher lead to any Hobbits receiving fewer education services than before the voucher program? In what case will the voucher lead Hobbits to spend exactly $1,000 more on education? In what circumstances will the voucher be treated just like a cash transfer?

Better off spending less on education and more on other goods when public option opens.

Charter Versus Private Schools (Vouchers)

Charter schools and vouchers both increase school choice, but... Lottery entrance to charter schools find positive results Lottery entrance to private schools find negative results The best charters tend to be nonprofit public schools open to all accountable to public authorities. The less "private" that school choice programs are, the better they seem to work

Impact of Out of School Inputs

Coleman Report (1966 ) Federal study: "Equality of Educational Opportunity" 650,000 students in the sample Segregated black and white schools received nearly equal funding, had different test scores Breakdown between family background and school influences was the core finding -Estimates 2/3 out of school, 1/3 in school factors -Teacher quality (educational attainment and experience) to be most important. -------------------------------------------------------- Hoxby (2001) uses regressions to apportion the explained variation in the students' math scores 𝑇𝑒𝑠𝑡 𝑆𝑐𝑜𝑟𝑒𝑠=𝑓(𝐹𝑎𝑚𝑖𝑙𝑦 )+𝑓(𝑆𝑐ℎ𝑜𝑜𝑙 )+𝑓(𝑁𝑒𝑖𝑔ℎ𝑏𝑜𝑟ℎ𝑜𝑜𝑑)+𝑒𝑟𝑟𝑜𝑟 Family variables: Parents education, family income, race, ethnicity, number of siblings; parental interest (attended school events/planned courses with child), books/calculator at home, family visiting a museum/library School input variables: Per-pupil spending, class size, teachers - min/max/average salary, % certified in their teaching area, % masters' degrees, years experience, books/computers per student, counselors per student Neighborhood variables: Census region residency, plus school district: mean income, income inequality, % below poverty line, % incomes above $50,000,% race, % high school degree, % college degree

Bias in Class Size

Compensatory - low-income students have smaller class sizes Income : + test scores, + class size → +ive bias If 𝛽_𝑐𝑙𝑎𝑠𝑠 negative then this would make class sizes look less important Complementary - high-income students have smaller class sizes Income : + test scores, - class size → -ive bias Would make class sizes look more important for academic achievement than they are.

School Choice Policies - Home schooling

Currently about 3% of kids are home schooled (up from 1.7% in 1999) Correlated with rise in online learning States differ markedly in their regulations concerning reporting requirements: -Require no notice of homes schooling by parents -8 States -Require curriculum approval, notification of test scores - 5 States Barriers Varies by state Economies of scale Choice Set Largest possible

Mechanisms causing perpetual gaps in performance

Early years: Importance of the home-learning environment -Reading books to children, computer at home Primary school: school choice, peer effects Secondary school: teenagers' and parents' expectations for higher education, access to material resources, anti-social behaviour Across childhood: parents' own cognitive abilities

The Market for Education (Economist View)

Education: Only 33% agreed that students would be better off if they all had access to vouchers to use at any private (or public) school of their choice Why? People can make informed choices only if they know a product's qualities and price e.g. why automakers are legally required to disclose their cars' fuel efficiency Relying on parents, not government, should be the sole judge of school quality may lead to bad results Require similar standards across school types

Hanushek Conclusion

Eric Hanushek "... Constantly rising costs and "quality" of the inputs of schools appear to be unmatched by improvement in the performance of students. It appears from the aggregate data that there is at best an ambiguous relationship and at worst a negative relationship between student performance and the inputs of schools. Such conclusions cannot, however, be made on the basis of aggregate data." (1986) So we need to apply econometric methods!

What is a duty-to-bargain law?

From textbook: A duty-to-bargain law states that employers have a legal duty to engage in collective bargaining in good faith with their employee's elected union of choice. From google: "Good-faith bargaining generally refers to the duty of the parties to meet and negotiate at reasonable times with willingness to reach agreement on matters within the scope of representation; however, neither party is required to make a concession or agree to any proposal." These were seen as pro union because they make it illegal for districts to refuse to bargain with teacher unions

Neighborhood Effect

Gibbons (2010) studies the impact of childhood neighbourhood status on educational attainment by age 33 Children of social tenants in the 1970s were influenced by the proportion of highly qualified adults in their neighbourhood Finds evidence of educational spillover effects from the community to the individual But this is only a small effect - nothing compared to the effect of parental background!

Education Production Function

Goal: Determine policies that are both technically efficient and allocative efficient -Relationship is key to determining efficiency of education producers -This relationship would give policy makers info on best ROI in ed spending

Demand Side - Informed Parents

Hastings and Weinstein (2008) Examined Middle School choices of parents All parents received information packs about schools' attainment In 2004 a simplified NCLB information pack was released They look at the changes in the application patterns If a school was labelled failing there was a 5.1% fall in applications to that school Takeaway: This implies that parents were not well informed before

Inflow - Teacher Selection

Improve the applicants: -Make the profession more attractive to everyone -Increase salary -Other Benefits Improve the selection process: -Standard measures for selection uninformative -Use of 'in the field measures' -Interview and a mock classroom interactions - Good predictor of future student gains -Over longer periods of time -Value-added measures of performance - Frist 2 years strong predictor of future student gains

US School Funding

In US, local control and of predominantly local financing of schools Highly localised resources - School district revenues highly dependent on local property tax Strict Catchment Areas - Limited choice of schools Thus living in a low income area -Property is less valuable -Schools will have fewer resources Hence, there are economic arguments for centralized mechanisms to reduce inequality in spending on schools.

School Choice: Supply Side (Choice Set)

In a competitive market - Average costs are non-decreasing in size A traditional education setting there are substantial economies of scale: -Large fixed administrative costs, start-up costs, specialized services -Average costs are lower for larger schools -Forms a barrier to entry for new schools -Optimal to have fewer larger schools - small choice set A traditional education setting parents have strong geographic preferences

School Choice: Supply Side (Barriers)

In a competitive market - No barriers to entry -There is entry into the market if 'firms' can cover their costs -Revenue from students at prevailing price must equal/exceed costs at given scale -Entrants will continue until economic profits become zero A traditional education setting - High barriers to entry District determine number of schools

School Choice (political views)

Increasing school choice is popular across the political spectrum Bush - No Child Left Behind - expanded school choice for students in failing schools. Obama - Race to the Top Initiative - Incentives for states to expand school choice programs for students in low-performing schools.

Major implications for the way in which the labor market for teachers should operate

Inflow - Those entering teaching - Selection/Training Methods (ability) Stock - Current Teachers - Nature of professional development (ability) / Incentives (effort) Outflow - Those leaving teaching - Improve retention/separation rate (ability/effort)

Input Based/Output Based Ex.

Input based policies - Class size, Sports, Mobile Phones Output based policies- School Choice, Measures of Teacher Output, Teacher Output Policies

Describe the five different types of school choice policies. For each one, explain what types of entities control the schooling option (i.e., public, private non-profit or private for-profit) as well as the way in which the schooling option is funded.

Inter/intra district open enrollment- public (controlled by any school in district or state) and public funded Charter schools- private nonprofit/for profit ran but publicly funded Vouchers - private nonprofit/for profit and publicly funded Private schools - private nonprofit/for profit and privately funded Home schooling - private nonprofit and privately funded

Sports in Schools

Last year 1,000,000+ students took part in high-school athletics This marks the 24th consecutive year of increase in high-school athletic participation Many observational studies show a positive correlation between high school athletes and: 1. Grades 2. Aspirations (Higher educational and occupational) 3. Time doing homework 4. Attitude towards school

How has the composition of teachers changed over time? What are the underlying causes of these changes?

Less likely to be drawn from elite colleges than they were 30 years ago. This is because of 1) the changing labor market opportunities for women and 2) the return to skill in most professions that require a college degree has grown, while teaching has exhibited increasing pay compression

Out of School Inputs - Glodhaber

Looks at the total variation in the student test scores within and across schools The paper does not try to assign what factors are important only where the variation happens 20% of total variance in test scores due to school effects both observable and unobservable

Impact of Private Schools (Vouchers) Why?

Louisiana's voucher program (5th largest) launched in New Orleans in 2008, expanded to include the entire state in 2012. Students from families with incomes below 250 percent of the federal poverty threshold are eligible for the voucher as long as they attend a public school the state has labeled as low-performing. Effectively an RCT Lottery winners have substantial achievement LOSSES Other studies Confirmed Why? -Transition year for students - But Charter schools have positive effect -Transition year for schools - Not used to standardised tests -Are test scores the important measurement -External validity - Are these schools representative of all private schools. Perhaps the participating schools were unusually bad and eager for revenue -Non-traditional students

Mobile Phones and Test Score

Method: Difference in Differences Banning Cell Phones in school improves high school test scores 6.90

Out of School Inputs - Combination of Methodology

More recent papers have tried combine both methods (Rasbash et al, 2010; Ainscow et al, 2010; Cook, 2013) By trying to account for the variation within school by controlling for characteristics of Individual Family Neighbourhood Leaving the impact of the school unspecified 𝑇𝑒𝑠𝑡 𝑆𝑐𝑜𝑟𝑒𝑠=𝑓(𝐼𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙 )+𝑓(𝐹𝑎𝑚𝑖𝑙𝑦 )+𝑓(𝑁𝑒𝑖𝑔ℎ𝑏𝑜𝑟ℎ𝑜𝑜𝑑)+𝑆𝑐ℎ𝑜𝑜𝑙+𝑒𝑟𝑟𝑜𝑟 19% of the variation in student test scores is due to school level factors The '80-20' rule

Impact of Sports on Test Scores

National Longitudinal Study of Adolescent Health 26,000 High schools from across the United States 132 High schools randomly selected Research supports that school athletics is associated with marginally better test scores

Out of school factors

Natural Ability Parental Investment Relative Age Neighbourhood Home Infrastructure Health Birth weight Genetics Health care Student Effort Food Insecurity Family Stress

Reconcile Findings (Hanushek)

Optimal allocation of resources means: -equating the marginal product per dollar of expenditure on all goods -Reaching PPF Suggests that it should be highly productive to give schools more money Schools may not be reaching PPF if: 1. Goals of the education system involve more than just maximizing student test scores. 2. Lobby groups may convince decision makers to allocate incorrectly 3. Little incentives due to low competition 4. Informational barriers

Coleman Report

Originator of Education Production Function -massive statistical base on schools and pupils in US -survey info on resources -distribution of resources by student background AIM: which of the various inputs into the educational process were the most important in determining the achievement of students

School Choice Policies - Vouchers

Parents are provided with private school tuition vouchers Milton Friedman (1955) "a stable and democratic society is impossible without widespread acceptance of some common set of values and without a minimum degree of literacy and knowledge on the part of most citizens," Mr. Friedman wrote, the government should pay for all children to go to school. Does not mean the government should run all the schools. Give parents vouchers to pay for "approved educational services" provided by private schools, with the government's role limited to "ensuring that the schools met certain minimum standards." Nationwide, 141,000 students use a voucher to attend a private school Barriers: -Quota on number of vouchers available for students Choice Set: -Families can spend more than the public allocation for schooling. -Increase choice for parental taste preferences f(.)

Skill Formation (Cuhna and Heckman) 2007

Persistent and substantial ability gaps across children from various socioeconomic groups emerge before they start school The family plays a powerful role in shaping these abilities through parental investments, and through choice of child environments Levels of child skills are highly correlated with family background factors like parental education and maternal ability Individuals use a production function to determine the relationship between inputs and the output of skill Parents make the investment decisions Outputs are the skills produced at each stage But there is self-productivity -skills produced at one stage raise the productivity of skills attained at later stages -Investments at one stage can lose their effects if not followed up in each period Because of this 'skill multiplier', early investment will yield the highest economic returns But investment must be kept up

Class Size as an Input (Mechanisms)

Potential Mechanisms: -Teachers have more time per student to spend on individualized instruction -Less likely to have a disruptive student in the class -Teachers may be able to tailor their lessons to the specific needs of the class -Potential high ability teachers may choose to become teachers

In terms of the education production function, what are the potential costs and benefits of giving principals more flexibility in determining how resources should be allocated? How do these costs and benefits relate to the relative desirability of resource, input and output based policies?

Potential benefits: principals have better sense of which inputs maximize their specific student's achievement level Potential costs: lack of goal setting/standardization

Under what conditions would the complete privatization of schools generate an efficient level of education investment?

Private schools with voucher system

Output Based School Policy: School Choice (Pro/Con)

Pro -Increasing school choice will inject more competitive forces into public education, and benefitting all students. Con - Education is not a competitive market, so moving public funding resources and students away from traditional public schools will harm the public education system more broadly,

Project STAR

RCT in TN 80 elem schools, 11,000 students cost: $12M Looked at effect size of smaller class sizes K-3rd grade It was statistically sig (via t-test)

Comparing Coefficients: Improve Test Scores

Reducing Class Size from 23 to 15 students improves kindergarten test scores by 4.83 -Small Class Impact 4.82, Standard deviation of outcome 26.7 Effect Size = Sports Participation improves students GPA by 0.04 Participate in 1 Sport Impact 0.04, Standard deviation of outcome 0.5 Effect Size = Banning Cell Phones in school improves high school test scores 6.90 Ban Phones Impact 6.90, Standard deviation of outcome 100 Effect Size = Having these estimates allow schools to establish which are the most effective But optimal policy depends on marginal benefits and marginal costs.

80-20 Rule Implication

Roughly 80% of the variation in children's school results is due to individual and family characteristics outside the school gates, with the remaining 20% due to what actually happens during school Implications : -Being in the very best household would put you in the 80th percentile of achievement -The impact of an average pupil going to the very best school instead of the would be an improvement in his or her results by one GCSE grade -Average student would go from a C to a B

Sports Participation and Absenteeism

Show that active competition decreases absences

Requirements for school choice to generate gains through competition For this to occur we require the education market to be competitive

Supply Side: The level of competition and the range of educational offerings in a local area -No barriers to entry or exit -A choice set Demand Side: The process by which students and their parents choose among educational options -Information: average grades, value added, f(.) -Preferences: School emphasis, learning style

Teachers

Teachers are currently of great policy interest Teachers salaries form the largest part of school expenditure For these lectures we are going to use a basic framework to simplify teachers decisions Objective: Educating pupils (academic ability, soft skills) Researchers have used many metrics, but we focus on academic achievement Very difficult to measure exactly Teacher Effectiveness determined by 1. Teacher Talent 2. Teacher Effort

Teacher Effectiveness Summary

Teachers have a large impact on student outcomes There are a range of ways to measure impact Characteristics, Test Scores, Observations, Surveys We can judge these metrics on Validity Stability Precision All measures have their problems The assumptions for VA rarely hold, however provide estimates very close to year on year experimental estimates, and are highly correlated with personal evaluations Using measures over a longer period of time reduces the instability Improving the implementation of the testing procedures would improve the precision of the estimates

What problems could be caused when using these measures for Performance Related Pay?

Teaching To The Test (TTTT) Actions taken by the teacher to improve high stakes results whilst ignoring low stakes Narrowing of the taught curriculum Coaching Cheating If test scores were the outcome we ultimately cared about then there would be no problem Therefore improving test design and testing procedures could reduce all these factors

National Assessment of Education Progress (NAEP)

Tests a nationally-representative set of students on a comprehensive set of subjects since the 1973 Students are tested at age 9, 13 and 17 These exams are taken intermittently The tests are designed to be comparable over time, knowledge and skills of American students. We focus on the mathematics and reading

Impact of Out of School Variables - Hoxby Paper

The Hoxby paper claims that 97.3% pupil outcomes explained by out of school factors Interpretation: An ideal family background would put the student in the 93percentile of pupil achievement, being in the best school would only increase test scores by 2.8percentiles What the paper is actually saying is - this is the proportion of explained variance, explained by family background variables. If these variables explain 40% of the variance. Then family background variables would only explain 38% (.97*0.4) of total student variance. The paper assumes these variables account for all family background/school inputs If the school variables are 'less good' at explaining than the family background ones This would bias this finding in favour of family background characteristics

Value Added Models - Precision

The precision depends upon 1) Number of observations 30-60% of the variation in measured teacher performance is due to sampling error Only a small number of observations per teacher per year 2) Improving the underlying measure Are increases in student test scores the correct metric in which to measure quality? Metrics would ideally capture student knowledge/understanding/creativity Timing of tests is also important

Relative Age Effects

There are significant gaps between August born and September born children at all levels August penalty is larger for those at the bottom of the ability distribution Evidence that the effects on test scores fade out Finding replicated in other settings (Bedard and Dhuey, 2006; Datar, 2006; Elder and Lubotsky, 2009; Puhani and Weber, 2007; Smith, 2009; Crawford, Dearden, and Meghir, 2007) But also find continuing effects into later life Higher university enrolment rates (Bedard and Dhuey, 2006) Higher adult wages (Fredriksson and Öckert, 2008; Kawaguchi, 2011)

School Choice Policies - Charter schools

They are privately-run but publicly-funded schools Enroll students regardless of the student's zoned school district Districts financing equal to the average cost per child in the district Charter schools are granted autonomy from many local regulations Highly varied - Ownership / Teaching style: 1,500 in 1999 to 5,700 in 2011 - Attendance increased from 0.7% to 4.2%. Barriers Quotas on number of charter schools that can locate in an area States set minimum standards for keeping charter status If replacing existing school, low start-up costs Choice Set Can select schools from out of district More Schools

Stock - Teacher Effort

To be an excellent teacher required talented and effort Policy makers have tried to design and implement mechanisms to increase effort Pay for performance Many potential issues: -What defined performance -Perverse effect of incentives -Design of performance scheme -Specific issues -Value added -Observations -Student surveys

3 Types of Education Policies to Increase Student Achievement

Total Resource Policies - these policies simply increase or decrease total school revenues, schools choose their expenditures so that the effect of spending an additional dollar on each input is equalized Input based - input policies alter a particular input Output based - these policies provide incentives for schools and teachers to increase output by setting targets

Explain the difference between a total resource policy, an input policy and an output policy. Give an example of one of each type of policy.

Total resource policy: These policies simply increase or decrease total school revenues -> schools choose their expenditures Input policy: Alter a particular policy/input Output policy: Provide incentives to teachers and staff to increase output by setting targets

Teacher Effectiveness Any metric can be assessed on three features to evaluate the usefulness

Validity - is the measure unbiased? Stability - is the measures consistent over time? Precision - is the measurement accurate? Ex. Drilling for Oil Validity - Are the number of barrels consistently under or over estimated? Stability - Are the number of barrels measured consistent over time? Precision - Can you get a precise estimate?

Teacher Policies Summary

Various metrics available with their costs and benefits as part of performance related pay -Value Added Measures - Objective, Predictive / Narrow, Noisy -Observations - Broad, hard to game / Non-Predictive, Political -Student Survey - Broad, Stable / Easy Classes, Grade Inflation Performance related pay may not be needed if teachers already acting optimally

LBJ Great Society - 3 Consequences

War on Poverty Elementary and Secondary Education Act (ESEA) Three major consequences 1. Federal funding to high poverty schools to close the achievement gap. 2. Forbid the notion of Federal Government setting a national curriculum 3. Set standards and achievements This was only about financing, not how to spend it. It was inputs based not outputs based, and assumed schools knew how to spend it. 1960's = $150 per K-12 students 2011 = $1600 per K-12 student

Teacher Impact

What does this mean? Highly Effective Teacher: A teacher one standard deviation above the average Having a very effective teacher as opposed to an average teacher is comparable to Reducing class size by ten in Year 5 (ages 9-10) Reducing class size by thirteen in Year 6 (ages 10-11) (Rivkin, Hanushek and Kain, 2005). An additional 25-45% of an average school year in maths performance (Aaronson, Barrow and Sander, 2007) For pupils from disadvantaged backgrounds the gains are larger "Students gain an extra year's worth of learning under very effective teachers compared to poorly performing teachers" (Hanushek, 1992).

Hanushek Critique What evidence does he use to support this claim?

a. Hanushek Critique states that there is little correlation between the amount schools spend on students and measured student academic outcomes b.Used: Time series evidence - large increases in per-student expenditures over time have not been met with gains in measured student achievement AND Education production function evidence - the research on education production functions has not found consistent evidence of a positive link between total school resources and student outcomes or between key inputs such as teacher salaries and student teacher ratios and student performance

7. Suppose a school's total revenue is $10,000,000 per year. The cost of hiring a teacher for a year is $50,000, and the cost of leasing a computer for a year is $5,000. Write the equation for the budget constraint. Draw the budget constraint. What is the relative price of teachers in terms of computers? How does the budget constraint show these relative prices? Can the school afford to hire 200 teachers and lease 10 computers?

a. 10000000=T(50000)+C(5000) b. c. 1 teacher = 10 computers The budget constraint shows these relative prices through the slope (2000/200) = 10 d. NO -->Would MP(t)=5000C + 50,000 and MP(c)=50,000t + 5000 ??? or is MP uncertain since the educ production fxn wasn't given

In 1985, then Governor Lamar Alexander and the Tennessee legislature implemented the Project STAR experiment. a. What was the goal of this experiment? b. Describe the experimental design c. Describe the findings

a. Random Control test to see the effect of smaller class size on test scores in TN b. Four-year experiment focused on students in the early elementary grades, starting in kindergarten and then moving to third grade A key to the research design is that both students and teachers were randomly assigned to the three types of classrooms. Critically, the randomization was done within each school. This assured that, on average, the characteristics of students and teachers across the different classroom types within each school were the same. Thus, within-school comparisons of student test scores across different class sizes yields the causal effect of class sizes on student achievement. 3 groups: 1. Control (regular class 22-25 students) 2. Treatment #1 (13-17 students) 3. Treatment #2 (regular class size with TA) Both teachers and students randomly assigned to the 3 types - randomization done w/in each school c. -Main findings suggest that smaller classes lead to higher student achievement -Overall, the results of Project STAR showed that assignment to a small class was associated with an effect size of about 0.2 -Estimates were larger in K and 1 st grade and then faded out, which led to criticism

Attending a less than excellent school _______ be compensated for.

can

Total Resource Policies (Hanushek Critique) What was the response to critique?

look at aggregate test scores, pretty damning on spending (spending has tripled but test scores have been flat) Black-white achievement gap has shrunk, test scores may not be the best thing to look at, better life outcomes, cherry-picking his research, he was counting all research equally → How we can resolve them: US has spent so much that it has diminishing marginal returns

Children from wealthier backgrounds are likely to attend schools with __________ classes

smaller

Define the term "education production function." 1. Name three potential inputs to the education production function that come from schools. 2. Name three potential inputs to the education production function that come from outside of school. 3. How does the fact that inputs come from the schooling environment and from other areas of children's lives complicate the study of how school-based inputs impact educational outcomes?

the process by which the outcomes of education, such as cognitive ability and knowledge are produced from the "raw" inputs 1. Class Size, Participation in Sports, Teacher Skill Level (effectiveness), Cell Phone Bans, Teaching materials 2. Family Stress, Natural Ability, Parental Background, Relative Age, Neighborhood 3. It is hard to completely control the variables in the study and eliminate all biases, to show causation. Also, not all parents are okay with putting their kids education at stake for the sake of a study, so there cannot be a completely randomized group if parents are intervening. (Parental involvement can be a bias in itself)