Research Class - Quantitative Data Part 1 2/6/21 lecture

8 criteria for evaluating inferential statistics:

(a) basic descriptive statistics are needed to evaluate the results of inferential statistics; (b) inferential analyses refer to statistical, not practical or clinical, significance; (c) inferential statistics do not indicate external validity; (d) inferential statistics do not indicate internal validity; (e) the results of inferential tests depend on the number of subjects; (f) appropriate statistical tests should be used; (g) the level of significance should be interpreted correctly; and (h) be cautious of statistical tests with small numbers of subjects in one or more groups or categories.

Criterion rated/predicted validity

-A test of whether a measure correlates with a known indicator or the behavior being measured -Deals with whether the assessment scores obtained for participants are related to a criterion outcome measure; predicative value -acquired through observation or experimentation -more empirical -manipulation of variables -Example: Do SAT scores predict college performance?

Advantages of quantitative data

-Conduct in-depth research: Since quantitative data can be statistically analyzed, it is highly likely that the research will be detailed. -Minimum bias: There are instances in research, where personal bias is involved which leads to incorrect results. Due to the numerical nature of quantitative data, the personal bias is reduced to a great extent. -Accurate results: As the results obtained are objective in nature, they are extremely accurate.

Internal validity

-Did the experimental manipulation really make a difference? -Research design must rule out all variables that might affect the dependent variables -Did the methodology control for variables that could affect the outcome or the dependent variable? -Factors that can affect are: History, Maturation, Attrition, Instrumentation, Testing Effects -your degree of confidence. the causal relationship being tested is trustworthy. It is not influenced by other factors

content validity

-Each item is judged for its applicability to the characteristic being measured -How well the content of a test reflects the subject matter from which conclusions will be drawn -How well does the test being administered accurately reflect the conclusions will be drawn? -you look at each item on a test & judge its applicability to the concept being measured -often judged by the researcher -more non empirical evidence -the adequacy with which the instrument samples the domain of measurement -construct a table of specifications

Quasi-experimental design/casual comparative

-Establish a cause-effect relationship between two or more variables -Researcher does not assign groups & does not manipulate the independent variable -Control groups are identified & exposed to the variable -Results are compared with results from groups not exposed to the variable -is not experimental because you are not using random groups true controlled groups -the controlled groups are kind of pre-exisitng

Definitions & benefits of quantitative data

-Examine the relationship between variables with the goal being to analyze and represent that relationship mathematically through statistical analysis -Focus on numbers, numerical data, measurement, control, manipulation, and experimentation -Quantitative data is defined as the value of data in the form of counts or numbers where each data-set has an unique numerical value associated with it. -Used to answer questions such as "How many?", "How often?", "How much?". -Standardized, objective, credible, limited bias -Numerical results can be displayed as graphs - easier to interpret -Data analysis is less time-consuming and can often be done using statistical software -Results can be generalized if the data are based on random samples and the sample size was sufficient -Data collection methods can be relatively quick, depending on the type of data being collected -Numerical quantitative data may be viewed as more credible and reliable, especially to policy makers, decision makers, and administrators -won't have as much bias because it is done statistically through a computer program -it can be verified using mathematical techniques

Examples of quantitative research questions

-How many new words do toddlers learn after one month of focused stimulation? -What percent decrease is seen in disruptive behaviors during treatment when a sensory activity is introduced before the session? -How often do patients with silent aspiration will eventually be treated for pneumonia?

Construct Validity

-Investigates the qualities a test measures -Deals with whether the assessment is measuring the true concept that is says it is measuring/correct construct (ability/skill) -does the test investigate the thing that it is really looking to investigate -acquired through observation or experimentation -more empirical -manipulation of variables

Measurement scales/variables

-Nominal level -Ordinal level -Interval level *used more in quantitative -Ratio level *used more in quantitative

Quantitative research goals

-Purpose/Goals: -Test theory -Establish facts and validate -Show relationships like cause and effect -Explain and predict -Statistically describe -Samples have to be carefully designed and chosen, or else their results can't be generalized

Limitations/Disadvantages of quantitative research

-Restricted information: Because quantitative data is not descriptive, it becomes difficult for researchers to make decisions based solely on the collected information. -Depends on question types: Bias in results is dependent on the question types included to collect quantitative data. The researcher's knowledge of questions and the objective of research are exceedingly important while collecting quantitative data -doesn't answer the why questions -can't tell you people's thoughts or perceptions

External validity

-To what extent can the results of a study be generalized to the population as a whole? -Representativeness? Does it represent the population that you need to be using this treatment with? Is the Practical/Clinical use is established? -The steps needed for external validity often decrease internal validity. The more real world your experiment is the less you are able to control all of your internal factors -Selection bias & practice effects can threaten EV -How does it apply to the real world? -people do field studies to go out there and understand the external validity of an experiment

Experimental Designs

-Use the scientific method to establish cause-effect relationship among a group of variables in a research study -Researcher makes an effort to control for all variables except the one being manipulated (the independent variable) -Effects of the independent variable on the dependent variable are collected and analyzed for a relationship

inferential statistics

-are used to compare groups -type of statistics used to make inferences about whether relationships observed in a sample are likely to occur in the larger population, that is, infer characteristics about a population based on data from a sample. -Statistical inference consists of two major approaches: estimating parameters and testing hypothesis. Parameter estimation is used to estimate a single parameter such as a mean. -Estimates can take two forms: point estimates or intervals estimation.

What is point estimation?

-calculated by dividing the observed values from the sample by the size of the sample, that is, a single statistic to estimate the population parameter. - is described as an educated guess on the basis of the sample data about the unknown value of the population.

Interval Data

-categories, rank order, equal spacing -has values of equal intervals that mean something. -Examples: a thermometer might have intervals of ten degrees; There is no true zero -additional examples include: temperature, IQ, SAT scores

Types

-counter -measurement of physical objects -sensory calculation -projection of data -quantification of qualitative entities

What is discrete data?

-data are described as having a finite number of possible values; space between values on a number line, thus values must be a whole number -Example: 100 questions on the test; # of correct out of a possible 100 Example: # of vehicles owned in a US household (e.g, 2 or 3 or 4) -example: rolling a dice, machines in a gym (they are all whole numbers, you can't have a fraction of a machine)

What is continuous data?

-data described as values that fall on a continuum; possible to have decimals or fractions -Example: height - 5'8" -Example: weight 123 lbs. -Example: distance - 14.3 miles -example: areas of lawn in square feet (you can have decimals or fractions when measuring a lawn)

Ratio Data

-exactly the same as the interval scale except that the zero on the scale means: does not exist. -There is a true and meaningful zero, 0,1,2,3,... -Examples: Age, Height, Weight, Number of children

Face Validity

-judged more by the researcher -more descriptive -more non empirical evidence -the appearance of the instrument, make the instrument appear appropriate

ordinal data

-qualitative data values with categories but it will be in a rank or ordering system -Shows position of one variable relative to another variable -Intervals are not equal -Uses non-parametric statistics -ex. mild, moderate, severe; High school class ranking: 1st, 9th, 87th...; Socioeconomic status: poor, middle class, rich; The Likert Scale: strongly disagree, disagree, neutral, agree, strongly agree; Level of Agreement: yes, maybe, no; Political Orientation: left, center, right.

Nominal Data

-qualitative data which consists of names, labels, or categories. -named, counted or labeled data -lowest level of measurement -classifies an event into a category -categories should be mutually exclusive -ex. male/female, smoker/non-smoker, typically developing/disfluent, Asian/Hispanic/Native American -equality, mode

Types of qualitative research designs

-quasi experimental design/casual comparative -experimental designs

Both external & internal validity can be improved by?

-randomization -choosing the correct participants -by using the correct statistics to analyze the data

What is the goal of inferential statistic?

-to discover some property or general pattern about a large group by studying a smaller group of people in the hopes that the results will generalize to the larger group. -So your smaller group is the sample and your larger group (that you want to learn about) is the population. -Another goal is to prove the null hypothesis incorrect -The logic says that if the two groups aren't the same, then they must be different. -inferential statistics uses hypothesis tests, confidence intervals, and regression analysis, a test you use might be a t-test

T/F Quantitative research can also be true experiments, pre-experimental, quasi-experiments or structured observations (e.g. cross-sectional studies)

True

T/F Quantitative research can take the form of surveys or interviews

True

What is projection of data?

a marketer will predict an increase in the sales after launching a new product with thorough analysis using algorithms.

What is interval estimation?

a range of numbers in which the population parameter falls considering a margin of error - often called the confidence interval.

Experimental can include?

between subjects designs and repeated measures

Types of data

discrete and continuous

In order to decrease your margin of error you must?

increase your sample size. Your margin of error is going to be larger if you have a smaller sample size

What is measurement of physical objects?

the HR executive carefully measures the size of each cubicle assigned to the newly joined employees.

What is counter?

the number of people who download a particular application from the App Store.

What is quantification of qualitative entities?

Identify numbers to qualitative information. For example, asking respondents of an online survey to share the likelihood of recommendation on a scale of 0-10.

What is sensory calculation?

Mechanism to naturally "sense" the measured parameters to create a constant source of information. For example, a digital camera converts electromagnetic information to a string of numerical data.

What are the weakest of all designs because there is no randomization or manipulation or use of control groups. It includes descriptive and correlational studies.

Non-experimental/observational