Strong Interest Inventory Certification Program Program Pre-Reading Assignment

Transcription

1 Strong Interest Inventory Certification Program Program Pre-Reading Assignment Strong Interest Inventory is a registered trademark of CPP, Inc. Readings: Strong Manual, Chapter 1, Strong User s Guide, Chapter 1, and specific readings referenced periodically Let s begin this section with an observation: Learning about measurement and statistics is usually not a favorite activity of those who are pursuing certification on the Strong! Apply your knowledge of Holland s RIASEC types to help answer why this might be? Often those who take this program have Artistic (A) and/or Social (S) in their Holland personality codes. For them, being creative and helping people is much more interesting than figuring out why an assessment scale works the way it does. Those with an Investigative (I) code may be more interested in this section. In order to use Strong assessment results to provide the best help possible to your students and clients, you need to be familiar with some basic statistical tools to make sense out of the numbers on the Strong Profile. These tools are used to interpret almost all career assessments. Once you learn them, you will probably use them a lot. You are likely to already know some or all of them. When you finish the Strong Certification Program you will be able to answer questions like these: A student asks, "Is my score a percentile?" An engineer in your company asks, "What percentage of the sample is included in the moderate range?" Your client asks, "If I take the Strong again, how likely is it that my results will be the same?" Your boss asks, "Why should we use this assessment that costs money rather than one we can get free on the Internet?"

2 For career development professionals who use career assessments... Understanding statistics is POWER! The power to help your students and clients get the most from their Strong results; The power to maintain credibility as a career professional; The power to justify the allocation of resources to programs that include career assessments. THE NUMBERS Many respondents will ask, "What do the numbers mean?" and "Where do they come from?" Many will just trust whatever you tell them about their results, and not pay much attention to the numbers at all. In either case, your understanding of the numbers is key to providing students and clients with accurate and meaningful information. The Strong is a self-report inventory, vs. a maximum performance test. Responses are scored on a five-point Likert-type scale of Strongly Like to Strongly Dislike--unlike performance tests which have right and wrong answers. This is probably why the word "test" can be kind of scary to students and clients, and why we always refer to the Strong as an "inventory." Likert items can be easily summed up to indicate the strength of a particular trait -- an interest in management, for instance. On the Strong, the Likert items are weighted this way: Strongly Like Like Indifferent Dislike Strongly Dislike 2 points 1 point 0 points -1 point -2 points So, if you respond "Strongly Like" to Being in charge of others, you get two points on the Management BIS. If you respond "Like," you get one point. If you say "Dislike," you lose one point on the Management BISs, and so on. The points that you are accumulating on the various scales add up to raw scores--numbers that are not seen on the Profile. All of the scores on the Strong Profile are standard scores--scores that have been rescaled so that data from different measuring systems can be easily compared.

3 MEASUREMENT TOOLS AND BASIC DESCRIPTIVE STATISTICS Any time you are dealing with groups of people or piles of data, it is useful to have some way of organizing the information. Descriptive statistics do that for you. The two types of descriptive statistics that we are concerned with on the Strong are: Measures of central tendency--the Strong uses a mean. The mean answers the question: where is the arithmetic middle of the group? Measures of variability--the Strong uses the standard deviation. The standard deviation answers the question: how spread out from the middle is the group of scores? Putting measurement tools and descriptive statistics in perspective, we have scores on the General Theme and Basic Interest scales that tell us where respondents are in relation to most people. Here is a picture of a group of people's scores on the Social General Occupational Theme (GOT), using the 2,250 members of the GRS (the General Representative Sample) on. GRS The General Representative Sample is a scientific sample of 2,250 working adults that is broadly representative of ethnic, racial, and occupational diversity. You can find a demographic profile of the GRS on pages of the Strong Manual. 68% You probably recognize this as a normal or bell curve. Most people (about 68%) will score in the middle of the scale--between about 40 and 60. In fact the average score of the General Representative Sample on the Strong scales is 50. This is why there is a big hump in the middle of the curve. Think of this as all these people piling up on top of each other. We refer to this as the mean score of the combined male and female GRS. Almost everyone (99.999%) will score between 20 and 80. We refer to this as the range of scores for the combined GRS.

4 Please turn to Table 3.2 in your Strong Manual: This table gives us the average/mean score for the RIASEC themes in the general population. Please look at the Social theme on this table: o The mean score for women is o The mean score for men is The SD (standard deviation) for men and women on each theme tells us how spread out the general population is on each RIASEC theme: o The General Occupational Theme (GOT) scale with the smallest spread in this table is Realistic: Women (8.42 standard score points). o The General Occupational Theme (GOT) scale with the largest spread is Conventional: Women (10.63 standard score points). o What this means is that women's scores don't vary as much--aren't spread out as much--on the Realistic GOT as they are on the Conventional GOT. Note: In all cases on the Strong, when we are referring to gender, we are referring to the identified gender of the respondent.

5 Standard Deviations To understand the concept of standard deviation, we will use the General Representative Sample, the general population group. Here are some things to keep in mind as you're "reading" standard deviations. You already did a little of this in Table 3.2 on the previous page: A large standard deviation means that the average distance of scores away from the mean is large; a small standard deviation means that the average distance is small. If scores cluster tightly together, there is not very much variability--everyone is scoring within a few points of each other, and the standard deviation will be small. If you have a large enough sample, 68% will fall in the middle two standard deviations. If your score is in the middle two standard deviations, it is considered to be just about average. This is what it looks like visually: A large standard deviation might look like this: the scores become more spread out. A small standard deviation might look like this: the scores are more similar to each other.

6 Correlation The statistical concept most often used to evaluate the reliability and validity of psychological assessments is correlation. This probably isn't the first time that you have run into correlations, but you may not have considered them statistically. Some of the things that we might talk about that involve correlations are: Calorie intake and weight Education and lifetime earnings Hours spent studying and academic performance Work satisfaction and productivity Age and wisdom Number of beers drunk during finals week and final exam grades Exercise and weight Driving speed and miles per gallon Carrots consumed and IQ These are all correlations, or co-relationships. They indicate a relationship between two things-- two variables-- that can be measured. The measurement can take one of three forms, referred to as the direction of the correlation: The correlation can be positive. Both variables move in the same direction: o As calorie intake increases, weight tends to increase. o The more education we attain, the higher our earnings tend to be. o The more hours we spend studying, the better our grades. o The more we like our work, the more productive we are. o As we get older, we get wiser. The correlation can be negative. The two variables move in opposite directions: o As alcohol consumption increases during finals week, grades go down. o As exercise increases, weight tends to decrease. o The faster we drive, the fewer MPG. The correlation can be zero. There is no relationship at all: o Eating carrots has nothing to do with IQ. But beware of statisticians! With a big enough sample, they can correlate ANYthing!

7 Consider these correlations and see if you can correctly identify their direction as positive or negative. They are a little tricky: Calorie intake and weight Answer: As calorie intake increases, weight increases: a positive correlation. Calorie intake and weight loss Answer: As calorie intake increases, weight loss decreases: a negative correlation. Hours worked and take-home pay for hourly workers Answer: The more hours worked, the bigger the paycheck: a positive correlation. AND: The less hours worked, the smaller the paycheck: also a positive correlation. Both variables go in the same direction. We will use correlations to evaluate the technical integrity of the Strong as we look at the relationships between: How a group of respondents score on the General Themes scale at one administration and how they score on the General Themes three months later (reliability) How a group of respondents score on half of the items on the Strong and how they score on the other half (reliability) General Occupational Themes and the Basic Interest Scales (validity) Basic Interest Scales and the Occupational Scales (validity) Basic Interest Scales and college majors (validity) Basic Interest Scales and occupational choice (validity) Each of these relationships can be evaluated using a correlation statistic, usually represented by an "r." You have already learned about the direction of the relationship: Positive correlations are positive numbers, for instance: r.80 Negative correlations are negative numbers, for instance: r -.80 These two correlations are equally strong. They just say different things. If we are talking about calorie intake and weight, for instance: r.80 says that as calorie intake increases, so does weight. And, if calorie intake decreases, so does weight. r -.80 says that as calorie intake increases, weight loss decreases. And, if calorie intake decreases, weight loss increases.

8 The plus or minus sign indicates the direction of the relationship between the two things we are considering. The number tells us the degree of the relationship. The value ranges from -1.0 to Correlations of -1.0 and +1.0 both represent perfect relationships. They never occur in real life. If they did, we would be able to predict the exact amount of weight loss from calorie intake, which of course we can't. There are many other variables to consider: gender, age, exercise, body type, etc. A correlation of 0.0 reflects no relationship at all, for instance between consumption of carrots and IQ scores. An important point about correlations: They do NOT indicate causality. Several years ago there was a story in the New York Times that reported, "a strong positive correlation between ice cream consumption and murders in New York City." What do you think? Does eating cream cause people to commit murder? It's not likely. Something else was going on in New York City that was related to both ice cream consumption and to murder. Any ideas? In the live program you will practice reading correlation statistics in the Strong Manual tables that illustrate the Strong's reliability and validity. There will be plenty of time for discussion and to ask questions about any statistics concepts that need clarification. We look forward to working with you as you learn about the Strong.