The purpose of this study was to examine the procedures for applying readability formulas to a standardized test and a modified version of the test. The personality construct of "locus of control" has been extensively used by investigators in a wide variety of situations and has occupied a central position in personality research for more than 15 years. Thus, different locus of control scales have been developed. In conducting research in an ABE environment, Levenson's Internal, Powerful Others, and Chance Scales (IPC) were considered to be a representative measure of the construct.

The IPC Scales were developed as a reconceptualization of Rotter's 1-E Scale, Levenson (1981) states that the multi-dimensional view of locus of control developed from questions about the validity of combining expectancies of fate, chance, and powerful others under the heading of external control. Individuals who believe in the influence of powerful others (one external orientation) will behave and think differently from individuals who feel the world is unordered and unpredictable (a second external dimension). In the former case, a potential for control over events exists.

The scales are composed of items adapted from Rotter's scale and items written specifically to measure beliefs about the operation of the three dimensions of control: beliefs in personal control (Internal Scale); powerful others (Powerful Others Scale); and chance or fate (Chance Scale).

The I Scale measure the extent to which people believe that they have control over their own lives (e.g. "When I make plans I am almost certain to make them work"); the P Scale deals with powerful others (e.g. "In order to have my plans work, I make sure that they fit in with the desires of people who have power over me"); and the C Scale is concerned with perceptions of chance control (e.g. "It's not wise for me to plan too far ahead because many things turn out to be a matter of good or bad luck"). (Lefcourt, 1981, p. 17)

Pretesting on 36 items included item analysis and correlations with the Marlowe-Crowne Social Desirability Scale (Crowne & Marlowe, 1964). The resulting IPC instrument has three eight- item subscales with a six-point Likert-type format. The test is administered to students as a single test having 24 items.

Internal consistency reliability estimates were only moderate. According to Levenson (1981), this is to be expected when the items sample a variety of situations. For an adult psychiatric sample of 115, alpha reliabilities were .51 for the I Scale, .72 for the P Scale, and .73 for the C Scale (Wallston, Wallston & Devellis, 1978). Levenson (1973) found similar estimates for a hospitalized psychiatric sample (.67, 182, and .79, respectively). Split-half reliabilities were .62, .66, and .64 for the IPC Scales. Test-retest reliabilities for a one-week and seven-week period were both in the .60 to .79 range (Lee, 1976; Levenson, 1973).

Levenson (1981) states that the validity of the scales has been demonstrated through convergent and discriminant methods (Campbell & Fiske, 1959) designed to show significant low order correlations with other measures of the general construct as well as a pattern of theoretically expected positive and negative relationships with other variables. However, low order correlations of the construct validation analysis are not clearly described.

The means and standard deviations of the IPC Scales for various adult populations such s psychiatric patients, cancer patients, the elderly, rural women, prisoners, and alcoholics have been reported (Levenson, 1981). In an evaluative study of suggestive-accelerative learning and teaching as a method of teaching vocational agriculture, a group of ninth graders were tested with the scales (Walters, 1977). However, one of the limiting characteristics of the instrument is that it has not been standardized on an adult basic education population. The sophistication of the wording may cause a percentage of the learners with low reading skills to not understand scale items.

There are five differences between the IPC Scales and the Rotter 1-E Scale (Levenson, 1981).

1. They are presented as Likert scales, instead of in a forced-choice format so that their three dimensions are most statistically independent of one another than are the two dimensions of Rotter's scale.

2. The IPC Scales make a personal-ideological distinction. All statements are phrased so as to pertain only in the person answering. They measure the degree to which an individual feels he or she has control over what happens, not what the person feels is the case for "people in general".

3. The items in the scales contain no wording that might imply modifiability of specific issues. Both factors of personal versus ideological control and system modifiability were found by Gurin et al. (1969) to be contaminating factors in Rotter's 1-E Scale.

4. The IPC Scales are constructed in such a way that there is a high degree of parallelism in every three-item set.

5. Correlations between items on the new scales and Marlow-Crowne Social Desirability Scale are negligible and nonsignificant. (Lefcourt, 1981, p. 18)

Because of these alleged measurement improvements, the IPC Scales were considered superior to the Rotter 1-E Scale for the purposes of the research with an ABE population. Although the Internal, Powerful Others, and Chance Scales (IPC) have been used with adult populations, further investigation of the scales' readability characteristics was deemed necessary. Since adult learners who have reading levels at the grade equivalencies from 5 to 12 enroll in community college retraining programs, it was necessary to use an instrument having a readability index of grade 5 equivalency, which was the assumed lowest reading level of the adult group. A number of readability principles were considered for use in modifying the IPC to make it consistent with the reading skills of the adult learner. The first step in assessing the readability of the IPC was to apply current readability formulas to the existing instrument. Three formulas were used: Fog Index, Fry Readability Graph, and the Flesch Formula. Laubach and Koschnick (1977) suggest the use of the Fog Index and the Fry Readability Graph for measuring the readability level of adult materials. The Fog Index was developed for use with the adult materials and yields scores equivalent to grade levels. Absolute values of the Fog Index, however, are often slightly higher than the scores derived from other formulas. Thus, the slightly higher score derived from the Fog Index provides greater assurance that readers will be able to cope easily and accurately with the assessed materials.

In contrast, the Fry Readability Graph is more useful to evaluators of materials than to writers preparing documents. It is routinely used by Literacy Volunteers of America. There are two notable limitations inherent in the formula: (a) in the word count, proper names and number are omitted for assessment purposes and (b) counting the number of syllables in each word tends to blur the perception of which are the difficult words. The Flesch (1974) formula was included in this study because it provides a human interest score derived from an assessment of the "personal words and sentences". Application of the Flesch formula yields two scores: reading and human interest. Each has a range of 0 to 100. The interpretation for reading ease is from unreadable to readable where readable means material that is understandable by people who have finished fourth grade and, therefore, who in the language of the Census are functionally literate (Flesch, 1974, p. 259).

BACK / NEXT PAGE