NCE Research and Program Evaluation B

If a distribution is bimodal, then there is a good chance that:

the researcher is working with two distinct populations. Imagine that you are plotting the average weights of adult men and women. In all probability, two distinct points of concentration would be evident on the curve.

A large study at a major university gave an experimental group of clients a new type of therapy that was intended to ameliorate test anxiety. The control group did not receive the new therapy. Neither the client nor the researchers knew which students received a new treatment. This was a:

double-blind study. A double-blind study goes one step beyond the single-blind version by making certain that the experimenter is also unaware of the subjects' status. In fact, in the double-blind situation persons assigned to rate or judge the subjects are often unaware of the hypothesis. This procedure helps eliminate confounding caused by "experimenter effects."

a CEEB score

expresses the abbreviation for the "College Entrance Examination Board" scores by the Educational Testing Service (ETS) of Princeton, New Jersey. This standard score is used for tests such as the GRE or the SAT. The scale ranges from 200 to 800 with a mean of 500. CEEB scores use a standard deviation of 100. Scores lower than 200 or above 800 are simply rated as endpoint scores. A score of 200 corresponds to 3 SD below the mean with 800 landing at a point 3 SD above the mean.

Test scores on an exam that fell below three standard deviations of the mean or above three standard deviations of the mean could be described as:

extreme. If you graph the situation you will note that these scores would be unusually high or very low.

In a new experiment, a counselor educator wants to ferret out the effects of more than one IV. She will use a ___ design

factorial. In a factorial experiment, several experimental variables are investigated and interactions can be noted. Factorial designs, therefore, include 2 or more IVs. Sometimes the IVs in a factorial design are called levels. The term "levels" does not connote hierarchy. You could have 2 levels of the IV such as "individual therapy" and "group therapy" but this does not mean that one is better than another.

In World War II the Air Force used stanine scores as a measurement. Stanine scores divide the distribution into 9 equal intervals with stanine 1 as the lowest ninth and 9 as the highest ninth. In this system five is the mean. Thus a Binet IQ score of 101 would fall in stanine:

five. Stanine is the contraction of the words "standard" and "nine." The mean or average score on the Binet is 100.

A counselor educator is teaching two separate classes in individual inventory. In the morning class the counselor educator has 53 students, and in the afternoon class she has 177 students. A statistician would expect that the range of scores on the test would be:

greater in the afternoon class in the morning class. The range generally increases with sample size.

an AB design

an AB or ABA time-series design is the simplest type of single-subject research and was initially a popularized by behavior modifiers in the 1960s and 1970s. The AB or ABA design relies on "continuous-measurement." A baseline is secured (A); intervention is implemented (B); and the outcome is examined via a new baseline (A) in the case of the ABA design. In order to improve the research process, an ABAB design can be utilized to better rule out extraneous variables. If the pattern for the 2nd AB administration mimics that of the 1st AB, then the chances increase that B (the intervention or so-called treatment) caused the changes rather than an extraneous variable. Some exams will refer to ABA or ABAB paradigms as "withdrawal designs." The rationale is that the behavior will move in the direction of the initial baseline each time the treatment is withdrawn if the treatment IV is responsible for the change. the ethical counselor must forgo using a withdrawal or reversal design (as they are sometimes called) if the removal of the treatment variable could prove harmful to the subject are those who come in contact with the individual. Here, a simple AB must suffice. Remember that when a researcher employs more than one target behavior, the term "multiple-baseline design" probably will be used in your exam.

a naturalistic observation

when clients are observed in a "natural" setting or situation

the John Henry Effect

(also called compensatory rivalry of a comparison group) is a threat to the internal validity of an experiment that occurs when subjects strive to prove that an experimental treatment that could threaten their livelihood really isn't all that effective. Say, for example, that counselor educators were asked to use computers as part of the teaching experience but were worried that the computers might ultimately take their jobs. The counselor educators in the comparison control group might purposely spend more time preparing their materials and give students more support than they normally would. One way for the researcher to handle this problem is to make observations before the experiment begins.

Z-scores (also called standard scores) are the same as standard deviations, thus a z-score of -2.5 means:

2.5 SD below the mean. this would be a T-score of 25. (This would also be a a CEEB score of 250. that is to say, you would take the CEEB mean of 500 and subtract 2 1/2 SD. since each SD on the CEEB scale is 100 you would subtract 250 from 500 which gives you a CEEB score of 250. It is conceivable that your exam could refer to a CEEB score as an ETS score.)

The range is a measure of variance and usually is calculated by determining the difference between the highest and the lowest score. Thus, on a test where the top score was a 93 and the lowest the score was a 33/100, the range would be:

61. The range is the simplest way to measure the spread of scores. Technically, statistics that measure the spread of scores are known as "measures of variability." The range is usually calculated by subtracting the lowest score from the highest score (e.g., 93-33=60). Some tests and statistics books define the range as the highest score minus the lowest score +1. if the test specifies the "inclusive range" then use the formula with +1. If not, go with the "exclusive range" formula which does not include it. My guess is that most counseling tests would give you either 60 or 61 as a choice (probably 60) but not both.

The variance is a measure of dispersion of scores around some measure of central tendency. The variance is the standard deviation squared. A popular IQ test has a standard deviation(SD) of 15. A counselor would expect that if the mean IQ score is 100, then:

68% of the people who take the test will score between 85 and 115. Statistically speaking 68.26% of the scores fall within plus or minus 1 SD of the mean. 95.44% of the scores fall within plus or minus 2 SD of the mean, and 99.74% of the scores fall within plus or minus 3 SD of the mean. Using this data, one could say that a person with an IQ score of 122 would fall within plus or minus 2 SD of the mean. Two SD would be IQs from 70 to 130 since 2 SD would be 30 IQ points. Please note that if everybody scored the same on the test then the SD would be zero. And SD, for example, of 1.8 has scores closer to the mean (i.e., not as spread out or scattered) than an SD of 2.8. The greater the SD, the greater is the spread.

Solomon four-group design created by psychologist Richard L. Solomon.

In this design, the researcher uses 2 control groups. Only one experimental group and one control group are pretested. The other control group and experimental group are merely post-tested. The genius of the design is that it lets the researcher know if results are influenced by pretesting. The 2 control groups as well as the 2 experimental groups can then be compared.

What type of data must be used with that Pearson "r" vs the Spearman rho?

Pearson "r," the most common correlation coefficient, uses I and R (interval and ratio data) as in "information and referral." Spearman rho ends in "o" as in ordinal.

Regardless of the shape, the ____ will always be the high point when the distribution is displayed graphically.

The mode will be the highest because it is the point where the most frequently occurring score falls.

mesokurtic

The normal Guassian curve is said to be mesokurtic since the peak is in the middle.

There are four basic measurement scales: the nominal, the ordinal, interval, and the ratio. The nominal scale is strictly a qualitative scale. It is the simplest type of scale. It is used to distinguish logically separated groups. The following illustrates the function of the nominal scale:

a DSM or ICD diagnostic category.the order of complexity of S.S. Stevens four types of measurement scales can be memorized by noting the French word "noir" meaning black. Parametric tests rely strictly on interval and ratio data, while nonparametric tests are designed only for nominal or ordinal information. The nominal scale is the most elementary as it does not provide "quantitative" (measurable) information. The nominal scale merely classifies, names, labels, or identifies by group. A nominal scale has no true zero point and does not indicate order. Other examples would be a street address, telephone number, political party affiliation, gender, brand of therapy, or number on a player's uniform. Adding, subtracting, multiplying, or dividing the aforementioned nominal categories would prove meaningless.

A bimodal distribution has 2 modes (i.e., most frequently occurring scores). Graphically, this looks roughly like:

a camel's back with 2 humps, for example a distribution of men and women's weights. When a curve exhibits more than 2 peaks it is known as a "multimodal" distribution. this can be contrasted to the curve with just a single peak (e.g., the normal curve) which is said to be "unimodal."

A sociogram is to a counseling group as a scattergram is to:

a correlation coefficient. A scattergram-- also known as a scatterplot--is a pictorial diagram or graph of two variables being correlated.

The ordinal scale rank-orders variables, though the relative distance between the elements is not always equal. An example of this would be:

a horse categorized as a second-place winner in a race. this is the second level of measurement. Nominal data does not rank-order the data like ordinal data. The rank does not indicate absolute differences. Thus, you could not say that the first, second, and third place horses were equidistant apart. The ordinal scale provides relative placement or standing but does not delineate absolute differences. Again, adding, subtracting, multiplying, or dividing is a no-no with this scale. Ordinal sounds like "order" so you should have no problem committing this scale to memory.

the range

a measure of variability, is the distance between the largest and the smallest scores. To compute the range you take the largest score minus the smallest score. The larger the range the greater the dispersion or spread of scores from the mean. Since the computation of the range is based solely on the computation of 2 scores, the variance and the standard deviation (the square root of the variance) are more stable statistics.

When a distribution of scores is not distributed normally statisticians call it:

a skewed distribution. In a skewed distribution the left and right side of the curve are not mirror images. In a skewed distribution the mean, median, and the mode fall at different points. In a normal curve they will fall at the same point.

participant observer model

a study in which the researcher actually participates in the study, while making observations about what transpired

In a normal curve the mean, the median, and the mode all fall precisely in the middle of the curve. From a graphical standpoint the so-called normal or Gaussian curve (named after the astronomer/mathematician K. F. Gauss) looks like:

a symmetrical bell. The normal curve is a theoretical notion often referred to as a "bell-shaped curve." The bell is symmetrical. most physical and psychological traits are normally distributed. In other words, if enough data are collected in regard to a given trait, and a frequency polygon is constructed, it will resemble the bell-shaped curve. NOTE: The 68-95-99.7 rule or empirical rule states that in a normal distribution 68% of the scores fall within plus/minus 1 standard deviation (SD) the meeting; 95% within 2 SDs of the mean; and 99.7% within 3 SDs of the mean. Almost all scores will fall between 3 SDs of the mean.

"Resentful Demoralization of the Comparison Group"

another control group phenomenon that threatens internal validity in research (also called "compensatory equalization"). Here, the comparison group lowers their performance or behaves in an inept manner because they have been denied the experimental treatment. When this occurs, the experimental group looks better than they should. If the comparison group deteriorates throughout the experiment while the experimental group does not, then demoralization could be noted. This could be measured via a pretest and a posttest.

A T-score is different from a z-score. A z-score is the same as the standard deviation. A T-score, however, has a mean of 50 with every 10 points landing at a standard deviation above or below the mean. Thus a T-score of 60 would equal +1 SD while a T-score of 40 would:

be -1 SD. Note that the T-score isn't as mathematically threatening since it is never expressed as a negative number. (-2 SD = a T-score of 30; a z-score of +2 = a T-score of 7-; a z-score of +1 = a T-score of 60.)

A good guess would be that if you would correlate the length of CACREP graduates' baby toes with their NCE scores the results would:

be close to 0.00. There is an absence of association here because as one variable changes the other variable varies randomly. The variation of one variable is most likely totally unrelated to the variation of the other.

a factorial design

can be used when there are 2 or more independent variables ( do not confuse with the term "factor analysis")

experimenter effects

can flaw an experiment because the experimenter might unconsciously communicate his/her intent or expectations to the subjects.

Dr. X discovered that the correlation between the therapists who hold NCC status and therapists who practice systematic desensitization is .90. A student who perused Dr. X's research told his fellow students that Dr. X had discovered that attaining NCC status causes therapists to become behaviorally oriented. The student is incorrect because:

correlation does not imply causal. Correlational research is quasi-experimental, and hence, it does not yield cause-effect data. A major research study, for example, might discover a very high correlation between the number of college students in a given geographical area and number of writing utensils owned. Yet it would certainly be misleading to conclude that owning a lot of writing utensils causes one to become a college student. Hint: When correlational data describe the nature of two variables, the term bivariate is utilized. If more than two variables are under scrutiny, then the term multivariate is used to describe the correlational paradigm.

skewed distributions

curves that are not symmetrical (i.e., those which are asymmetrical).

Experimental is to cause and effect as correlational is to:

degree of relationship. A correlation coefficient is a descriptive statistic which indicates the degree of "linear relationship" between 2 variables. Statisticians use the phrase "linear relationship" to indicate that when a perfect relationship exists (i.e., a correlation of 1.0 or -1.0) and it is craft, a straight line is formed. the Pearson Product-Moment correlation "r" is used for interval or ratio data while the Spearman rho correlation is used for ordinal data. Correlational research is not experimental and hence does not imply causality.

A distribution with class intervals can be graphically displayed via a bar graph also called a:

histogram. Most bar graphs are drawn in a vertical fashion. When the bars are drawn horizontally it is sometimes called a "horizontal bar chart." A "double-barred histogram" can be used to compare two distributions of scores such as pre-and posttest scores.

If an experiment can be replicated by others with almost identical findings, then the experiment:

is said to be reliable. The term "reliability" in the social sciences is also used in regard to testing to indicate consistency and measurement.

This would most likely yield a perfect correlation of 1.00:

length in inches and length in centimeters. In the real world, correlations may be strong (e.g., height and weight), yet they are rarely 1.00. Correlation is concerned with what statisticians call "covariation." When two variables vary together statisticians say the variables "covary positively," and when one increases while the other decreases they are said to "covary negatively."

The most useful measure of central tendency is the:

mean, often abbreviated by an X with a bar over it. in everyday life when we use the word "average" we are referring to the "mean." Perhaps this is because in most instances it is the most useful of the 3 measures of central tendency. Nevertheless, if a distribution is plagued with extreme scores than the "median" is the statistic of choice. The median is best for skewed distributions.

In a basic curve or so-called frequency polygon the point of maximum concentration is the:

mode; the score that appears the most (the point of maximum concentration)

Billy received an 82 on his college math final. This is Billy's raw score on the test. A raw score simply refers to the number of items correctly answered. A raw score is expressed in the units by which it was originally obtained. The raw score is not altered mathematically. Billy's raw score indicates that:

more information is obviously necessary. The fact that Billy scored an 82 tells you next to nothing. Raw data is like a raw piece of meat; it is uncooked and nothing has been done to it. How many questions are on the test? Well, you don't know, do you? So, you couldn't choose "he answered 82% correctly" since you don't have enough information to figure out the percentage. The question doesn't specify this critical fact. You see, if Billy scored an 82 on a task with 82 questions, then he had a perfect score. If, however, the exam had several thousand items, his score may not have been all that high. I say "may not have been all that high," since a raw score of 82 might have been the highest score if anybody tested. You would need a "transformed score" or "standard score" such as "his percentile rank is 82" to make this determination. NOTE: the benefit of standard scores such as percentiles, t-scores, z-scores, stanines, or standard deviations over raw scores, is that a standard score allows you to analyze the data in relation to the properties of the normal bell shaped curve.

An IQ score on an IQ test which was three standard deviations above the mean would be:

near the genius level. Think of it this way. Over 99% of the population will score between plus or minus 3 SD of the mean. Now that would be a very high IQ score; 145 on the WAIS-III to be exact. Lewis M. Terman, a pioneer in the study of intelligence, classified any children with IQs over 140 as "geniuses."

Aine of the world's finest counselor educators are given an elementary exam on counseling theory. The distribution of scores would most likely be:

negatively skewed. We would expect these folks to score really high; and thus, the right side of the curve would be packed with high values. This gives you a long tail that points to the left, which indicates a negative skew. The tail points you in the direction of the correct answer.

single case investigation

often called "ideographic studies" or "single-subject designs." The original case study methodology was popularized by Freud, though needless to say, unlike the behaviorists, Freud did not rely on numerical baseline measures. Case studies are often misleading because the results are not necessarily generalizable.

The standard deviation is the square root of the variance. A z-score of +1 would be the same as:

one standard deviation above the mean. Z-scores are the same as standard deviations! In fact, z-scores are often called standard scores. A z-score is the most elementary type of standard score. It is possible your exam will refer to it merely as a standard score. A z-score of plus one or 1 SD would include about 34% of the cases in a normal population. For those with a fear of negative integers, the normal distribution also can be described using T-scores, sometimes called "transformed scores." the T-score uses a mean of 50 with each SD as 10. Hence, a z-score of -1.0 would be a T-score of 40. A z-score of -1.5 would be a T-score of 35 and so on. If double-digit figures intimidate you, then you might want to analyze the normal distribution using a "stanine" score which divides the distribution into nine equal parts with 1 the lowest and 9 the highest portion of the curve.

The y axis is used to plot the frequency of the DVs. The y axis is also known as the:

ordinate. The ordinate plots the DV or experimental data.

If a group of first semester graduate students in counseling took the NCE exam, a distribution of scores would be:

positively skewed. Try to imagine this in your mind or roughly graph it on scratch paper. First semester students would probably have little information on the more advanced points of counseling; and thus, we would expect them to score very poorly. Hence, most of the scores would fall on the left or the low side of the distribution. Graphically then, the "tail" of the distribution would point to the right or the positive side. The tail indicates whether the distribution is positively or negatively skewed.

the harmonic mean

refers to a central tendency statistic that is the reciprocal of the arithmetic mean of the reciprocals of the set of values. For example, your exam asks you to calculate the harmonic mean for 3 scores of: 2, 2, and 4. 1st you would convert them to reciprocals: 1/2, 1/2, and 1/4. The arithmetic mean then is 1/2 + 1/2 + 1/4 = 1 1/4 or 1.25/3 = .4166. the reciprocal of this gives you a harmonic mean of 1/.4166. The statistic has limited usage; however, it is occasionally called for if measurements were not made on an appropriate scale (e.g., data revealed the number of behaviors per hour, when the number of minutes per behavior would be more useful). The harmonic cannot be utilized with negative numbers or if the data include a score of 0.

the observer effect

refers to a situation in which the person observing in a participant-observer research study influences/alters the situation.

In a new study the clients do not know whether they are receiving an experimental treatment for depression or whether they are simply part of the control group. This is, nevertheless, known to the researcher. Thus, this is a:

single-blind study. In the single-blind study the subject would not know whether s/he is a member of the control group or the experimental group. This strategy helps eliminate "demand characteristics" which are cues or features of the study which suggest a desired outcome. In other words a subject can manipulate and confound an experiment by purposely trying to confirm or disprove the experimental hypothesis. Let us say that in the referenced experiment a subject is fond of the researcher. And let us further assume that a score on a standardized depression inventory will be used as the DV. Our subject might purposely answer the questions as if he is less depressed than he really is. A subject who disliked the researcher could present himself as even more depressed.

Behaviorists often utilize N=1, which is called intensive experimental design. The first step in this approach would be to:

take a baseline measure. "N," or the number of persons being studied, is one. This is a "case study" of one approach. This method is popular with behaviorists who seek overt (measurable) behavioral changes. The client's dysfunctional behavior is measured (this is called a baseline measure), the treatment is implemented, and then the behavior is measured once again (i.e., another baseline is computed). Exams sometimes delineate this paradigm using uppercase A's and B's and C's such that A's signify baselines, B's intervention implementation, and C's a 2nd or alternative form of intervention.

The interval scale has numbers scaled at equal distances but has no absolute zero point. Most tests used in school fall into this category. You can add and subtract using interval scales but cannot multiply or divide. An example of this would be:

that an IQ of 70 is 70 points below an IQ of 140, yet a counselor could not assert that a client with an IQ of 40 is twice as intelligent as a client with an IQ of 70. since the intervals are the same, the amount of difference can be stipulated (e.g., 3 IQ points). Using this scale, distances between each number are equal yet it is unclear how far each number is from zero. Division is not permissible inasmuch as division assumes an absolute zero. (If you had an absolute zero then you could in fact assert that a person with an IQ of 140 would be twice as smart as someone with an IQ of 70. But of course, zero on an IQ test does not equal zero knowledge; hence, IQ tests provide interval measurement.)

The X axis is used to plot the IV scores. The x axis is also known as:

the abscissa. Again, it is the horizontal axis which plots the IV - the factor manipulated via the experimenter.

the mean

the arithmetic average

From a mathematical standpoint, the mean is merely the sum of the scores divided by the number of scores. The meeting is misleading when:

the distribution is skewed and/or there are extreme scores

In a career counseling session an electrical engineer mentions 3 jobs he has held. The 1st paid $10 per hour, the 2nd paid $30 per hour, and the 3rd paid a higher rate of $50 per hour. The counselor responds that the client is averaging $30 per hour. The counselor is using:

the mean

the median

the middle score in the distribution of scores

The median is:

the middle score when the data are arranged from highest to lowest. In studies measuring variables with extreme scores (e.g., family size or income), the median would be the best statistic. Rank the scores from lowest to the highest to find the middle score. If there are 2 middle scores, take the average of the 2 ( this may legitimately yield a fraction).

The most common measures of central tendency are the mean, the median, and the mode. The mode is:

the most frequently occurring score and the least important measure of central tendency. The mode is the highest or maximum point of concentration, the score that occurs the most. The modal score is the highest point on the curve. Hence, a test might tell you that a population of schizophrenics consists of 400 Caucasians, 60 Asian-Americans, and 100 African-Americans and ask you to pick out the so-called modal category, rather than the middle score.in this case the highest value is held by the Caucasian population.

ethological observation

the observation of animals

A platykurtic distribution would look approximately like:

the upper half of a hot dog lying on its side over the abscissa. If you see the word "kurtosis" on your exam, it refers to the peakedness of a frequency distribution. A "platykurtic" distribution is flatter and more spread out than the normal curve. This is easy to remember if you consider that "plat" sounds rather like "flat." When a curve is very tall, thin, and peaked it is considered "leptokurtic."

When a horizontal line is drawn under a frequency distribution it is known as:

the x axis. When the graphically representing data, the "x axis" (also called the abscissa) is used to plot the independent variable. The x axis is the horizontal axis. The "y axis" (also called ordinate) is the vertical axis which is used as a scale for the dependent variable.