Sample Lab
Sample Lab Exercise
This sample is not a complete lab, and is designed to illustrate how you mught approach answers to questions for the module 2 lab.
The attached article (Bach and Hayes, 2002) was co-authored by your instructor and will be used to illustrate your lab assignment since it is old enough that (1) your instructor is very familiar with its content (2) she is very familiar with all of its strengths and weaknesses (3) it is old enough such that it can be viewed by all of you without violating copyright laws!..
The research was a study comparing the effectiveness of a new type of therapy in the treatment of persons with symptoms of psychosis..
Notice that most (but not necessarily all) of the answers to the lab questions will be found in the methods section of the article.
In your lab you will have to consider:
1. From the article, identify and name:
- Two constructs
- Two operational Defintions
- The type of measure (e.g., nominal, ordinal, etc.) for two of the measures mentioned in the article, and the mode of each of the same two measures (e.g., self-report, biological, etc.)
2. write a sentence on the reliability and validity of the two measures (e.g., is there no information about their reliability, is it excellent or poor, what kind of validity info is provided?)
3. Idenitify if there are any potential ethical issues raised by the study
4. search the literature and find one different types and/or modes of meausrement for each of the measures you reported on in #1. For example, if your original article measured aggression using a self-report scale that was ordinal, you might find a measure that is observational, or uses a nominal scale.
5. For both measures write a few sentences on which measure - the one in the original aricle, or in one you report on in question 4 - you think is best and why you think it is best (e.g., better reliability, self report so easier to administer than biological, one has more info available on its validity.)
So...when considering my research report (Bach and Hayes, 2002)
(1) Psychosis is a construct, so it must be measured operationally rather than directly. In the psychosis study, ‘psychosis’ is operationalized as “reporting delusions or hallucinations at the time of admission to the hospital.” It could have been operationalized other ways, for example, by the score on a measure of psychological symptoms, or by having a certain diagnosis such as schizophrenia. (for more info on constructs, see p. 74-75 in your textbook and for operational definitions review pp. 75-77)
It is a self-report measure because the participants reported whether or not they were experiencing these symptoms. (see pp. 93 – 95).
(1) The rating of 'distress' is an ordinal measure because the ratings are organized in a sequence, but unlike an interval scale we can’t really say that a ranking of 4 means that the participant is “twice as distressed” as someone with a rating of 2 (see pp. 89 – 93 on scales of measurement).
(5) BTW, there was a potential floor effect to this measure – when people recover, they are usually less distressed, so at follow-up many of them reported a distress rating of zero (see pp. 96- 102 on ‘other aspects of measurement).
(1) Successful treatment is operationalized as days out of the hospital following treatment. This is a behavioral measure because the investigators observed whether or not the participant was hospitalized over a period of four months (behavioral measures p. 95).
(1) Days to hospitalization is an interval measure because there is no true ‘zero’ (if it were zero days to hospitalization, that would mean that the person never left the hospital in the first place, and would not make sense since the measurement began after they left the hospital), and it is interval rather than nominal because someone who stayed out of the hospital 20 days remained out twice as long as someone hospitalized after 10 days (scales of measurement p. 89-91).
(2) The authors created the distress measure themselves; this is not ideal, because it means there was no information about its reliability or validity (see pp. 77 – 87 on validity and reliability). But at least they could say that the measure has ‘face validity’ because it looks like a reasonable measure of symptom severity (see face validity on p. 78). The lack of reliability and validity info is a weakness…so why did they do it? (Answer below)
(2) There is some evidence for the reliability and validity of days to hospitalization because it has been used in other studies and is commonly associated with seriousness of illness and symptoms. It has high inter-rater reliability and good predictive validity.
(3) One potential ethical issue was informed consent. People experiencing symptoms of psychosis are not always mentally able to make good decisions. The author addressed this by only including participants if they did not have a legal guardian, and by not inviting them to participate in the study until they were stable enough to understand the questions being asked of them (informed consent p. 117-119).
(4) psychosis could be measured using the Brief psychiatric rating scale, which is on an ordianal scale, or by using the Peter's Delusional inventory whcih is on an interval scale.
(4) successful treatment could be measured by the Quality of Life Inventory, which is on an interval scale, or by change in Breif Psychiatric Rating Scale score, which is on an ordinal scale.
NOTE: I didn't cite my souce here for answeringquestion 4, and for your lab you will have to
(5) Answer to question above about whether the distress rating being a 'good' measure: One reason researchers sometimes create their own measures even when they lack data on reliability and validity is that they are studying a new phenomenon or problem and there are no reliable valid measures available. The research has to start somewhere! As others replicate the study the info about reliability and validity will grow. The measure of distress is not so good.
(5)The measure of days to hospitalization is widely used in research on treatment outcomes for psychosis, but quality of life is also a good measure because it looks at not only symptoms, but also how well the person is doing in many domains of life.