Homework answers / question archive / HPS104 Foundations of Psychological Science   Assessment Task 1 – Research Designs (30%) WORKSHEET Due Monday 8pm of teaching week 7 USE THIS WORKSHEET TO MAKE NOTES AND RECORD YOUR ANSWERS

HPS104 Foundations of Psychological Science   Assessment Task 1 – Research Designs (30%) WORKSHEET Due Monday 8pm of teaching week 7 USE THIS WORKSHEET TO MAKE NOTES AND RECORD YOUR ANSWERS

Psychology

Share With

---

AIM OF AT1

AT1 is designed to teach and assess primarily the following unit learning objectives:

1. Explain psychological phenomena: Describe the empirical and logical bases of psychological science
2. Test the validity of psychological explanations: Formulate a research question and select an appropriate research design.
3. Conduct research into psychological phenomena: Implement a research design in a manner that is valid, ethical and feasible.

INSTRUCTIONS

The worksheet. This worksheet contains the instructions for the assignment along with the questions you need to answer. Use this worksheet to record your ideas, working out/rationale, and answers to each of the questions and tasks. You can complete the worksheet in your own time and/or in the relevant Tutorial/Seminar. You do not submit the worksheet.

The weblink. After you complete the worksheet, please transcribe your responses to the weblink provided in the Assignments folder in Content (Cloud). Uploading your responses will allow me to process your answers and provide you with feedback more quickly, allowing you to resubmit later that same week if you achieve less than a credit grade for it.

Resubmitting. Once you receive your assignment feedback and mark via email you have the option of revising and resubmitting. You can resubmit using the same weblink you used for your original submission, and you can continue to resubmit up until the day of the exam. Note that you can only improve your mark up to 18/30 (i.e., 60%) for the assignment, but also that you will not lose marks on resubmitting (i.e., if your resubmitted mark is worse than your previous mark, you will keep your previous mark).

 

 

 

 

 

A TRUE EXPERIMENT

Experiments don’t get simpler than this! All we need is to recruit a group of willing and able doctors who are unaware of the full/true purpose of our study and randomly assign each of them either to receive the treatment or control version of the vignette.

 

QUESTION 1

[Each correct answer is worth 1/3^rd of a mark. You can earn a total of TWO marks for this question]

Please answer the following questions about each variable in your true experiment:

Variable name	Is it an independent (IV) or dependent (DV) variable in your study?	What measurement scale is used for this variable? (nominal, ordinal, interval, ratio)	Is this variable discrete or continuous?
Vignette type (treatment v control)	[   ]	[   ]	[   ]
Diagnosis rating (final slider position: 0-10)	[   ]	[   ]	[   ]

 

QUESTION 2

[Each correct answer is worth 1/2^th a mark. You can earn a total of ONE mark for this question]

How many levels are there of your independent variable? ______

How many conditions are there in your study? ______

 

 

 

QUESTION 3

[Each correct answer is worth 1/2^th a mark. You can earn a total of TWO marks for this question. Note, however, that your ranking of each issue will not be worth marks]

Identify whether each of the following is primarily important for internal validity or external validity. After you've categorised each feature into the appropriate category (internal v external), if you like, you can try to rank it in order of importance relative to the others within that same category (where 1=most important). Note that correct ranking isn’t worth marks.

 

INTERNAL VALIDITY	EXTERNAL VALIDITY	RANK*	ISSUE
[   ]	[   ]	[   ]	Ensuring the vignette received by doctors – treatment versus control – is randomly determined.
[   ]	[   ]	[   ]	Ensuring that the descriptions used in the treatment and control vignettes are identical aside from whether or not they mention a pre-existing mental illness diagnosis.
[   ]	[   ]	[   ]	Ensuring that doctors are from diverse backgrounds, ethnicities, gender, age, etc.
[   ]	[   ]	[   ]	Including diverse mental illnesses (not just schizophrenia) in the treatment vignettes.

*Remember, the rank isn’t worth marks.

 

HYPOTHESES

Before we start the project, before we collect any data in the project, we need to come up with two hypotheses: The null hypothesis and the alternative hypothesis.

Ho:         The null hypothesis specifies what the results of your study should look like if DO does not exist.

Ha:         The alternative hypothesis specifies what the results of your study should look like if DO exists and influences doctors’ diagnoses.

These hypotheses are worded such that they represent a direct contradiction to each other. Only one can be right, not both, and there is no middle ground between them. We want to decide, based on the results we eventually obtain, whether we should fail to reject the null hypothesis (there is not enough evidence to infer that there is an effect) or reject it in favour of the alternative hypothesis (there is enough evidence to infer that there is an effect).

Both hypotheses need to be simple, concise, relevant to the issue under investigation, consistent with the design employed (a true experiment), and need to include mention of all variables (and their levels!) included in the design.

 

 

 

QUESTION 4

[Only one hypothesis is correct – ranking this hypothesis at #1 is required to earn ONE mark for this question; how you rank the other hypotheses will not make a difference to your mark]

Below is a list of hypotheses. Identify the best NULL HYPOTHESIS by ranking it at #1 (the other rankings do not matter).

 

NULL HYPOTHESIS	RANK
HYPOTHESIS 1. Diagnosis ratings will be different for treatment and control vignettes.	[   ]
HYPOTHESIS 2. Diagnosis ratings will be equally low for treatment and control vignettes.	[   ]
HYPOTHESIS 3. Diagnosis ratings will be equally high for treatment and control vignettes.	[   ]
HYPOTHESIS 4. Diagnosis ratings will be the same for treatment and control vignettes.	[   ]

 

QUESTION 5

[Only one hypothesis is correct– ranking this hypothesis at #1 is required to earn ONE mark for this question]

Below is a list of hypotheses. Identify the best ALTERNATIVE HYPOTHESIS by ranking it at #1 (the other rankings do not matter).

 

ALTERNATIVE HYPOTHESIS	RANK
HYPOTHESIS 1. Diagnosis ratings will be the same for treatment and control vignettes.	[   ]
HYPOTHESIS 2. Diagnosis ratings will be different for treatment and control vignettes.	[   ]
HYPOTHESIS 3. Treatment vignettes will receive lower diagnosis ratings than control vignettes.	[   ]
HYPOTHESIS 4. Treatment vignettes will receive higher diagnosis ratings than control vignettes.	[   ]

Hint: Remember, diagnostic overshadowing refers to discriminatory medical treatment against people with a mental illness. For example, their symptoms may not be taken as seriously.

 

 

 

A QUASI-EXPERIMENT

Our true experiment only examined medical decision-making by doctors (general practitioners). Can we really extrapolate from their responses to medical professionals more generally? To find out we repeat the study on three groups: Doctors (as in the original study), nurses, and pathologists.

 

QUESTION 6

[All answers need to be correct in order to earn ONE mark for this question]

How would you describe your new study? (select only one option)

[   ]             a true experiment with two independent variables

[   ]             a person by treatment quasi-experiment

[   ]             a natural experiment

 

QUESTION 7

[Each correct answer is worth 1/2^th a mark. You can earn a total of ONE mark for this question]

How many IVs are there in this new study? ______

How many conditions are there now? ______

 

QUESTION 8

[All answers need to be correct to earn ONE mark for this question]

What is the most important thing you are doing by including practitioner type as a variable? Select only one option.

[   ]             improving internal validity.

[   ]             improving external validity.

 

 

 

QUESTION 9

[Each correct answer is worth 1/2^th a mark. You can earn a total of THREE marks for this question]

Identify whether each of the following statements is primarily relevant for internal validity or external validity. There is no need to rank these in terms of importance (but you can if you want to!).

INTERNAL VALIDITY	EXTERNAL VALIDITY
[   ]	[   ]	Attrition in the study is higher than I would have liked (it is 15%).
[   ]	[   ]	Attrition is higher in the treatment groups compared to the control groups
[   ]	[   ]	My random assignment method does not work as planned – I have accidentally placed all the nurses in the treatment group, all the doctors in the control group, and split the pathologists between the groups.
[   ]	[   ]	Several participants withdraw from the study when they discover that we will asking them to respond to made-up vignettes.
[   ]	[   ]	I have more trouble recruiting pathologists than nurses (i.e., my study has more nurses in it than pathologists).
[   ]	[   ]	The results of a different vignette study are completely different from mine. I notice that in their study they used major depression in their treatment vignette rather than schizophrenia.

 

A NATURAL EXPERIMENT

You discover that Clinic X typically withholds unnecessary information about patients’ mental health from doctors whereas Clinic Y, like most clinics, does not do this – their doctors receive patients’ entire medical history including mental health information. Could this be an opportunity for a natural experiment?

QUESTION 10

[Each correct answer is worth 1/2^th a mark. You can earn a total of TWO mark for this question]

Identify whether each of the following is primarily important for internal validity or external validity in the context of your natural experiment. There is no need to rank these in terms of importance.

 

 

 

INTERNAL VALIDITY	EXTERNAL VALIDITY
[   ]	[   ]	You discover that both clinics employ almost no foreign-trained doctors.
[   ]	[   ]	You discover that the two clinics have different types of patients. Clinic X tends to have more patients with mental health problems than Clinic Y.
[   ]	[   ]	You discover that Clinic X has more experienced doctors than Clinic Y.
[   ]	[   ]	You discover that Clinic X made its decision to withhold patients’ mental health history because its doctors had very discriminatory attitudes about mental illness.

 

 

 

PLANNING YOUR STUDY AND OPERATIONALISING YOUR VARIABLES

We will measure previous experience of mental illness stigma in the medical context using a newly developed survey instrument called the Concerns with Diagnostic Overshadowing Scale (CDOS).

 

The CDOS can be used to measure the mental illness stigma a PWMI has experienced in the medical context (i.e., the DO they have experienced) along a continuum (0-10) or as a discrete (yes/no) variable.

 

We will measure personal delays in health-seeking using the medical delays scale (MDS). Participants are basically asked whether they would delay seeking medical help for a physical illness if symptoms arose. This measure can be used either as a measure of the extent to which they would delay seeking medical assistance (i.e., along a continuum) or as a discrete (yes/no) variable.

 

We can now operationalise our research question in very general language as follows: Is CDOS score relevant to MDS score? We could depict this as a direct path: CDOS à MDS

 

 

IDENTIFYING YOUR DESIGN OPTIONS

Your new research question is: Is mental illness stigma associated with personal medical delays?

 

Notice how we are avoiding causal terms in describing our research question – we are, after all, using observational not experimental designs here. And among the observational designs there are several options available to us. Let’s go through some of these and see if you can name the design based on a description of its implementation.

QUESTION 11

[Each correct answer is worth ½ a mark. You can earn a total of FIVE marks for this question]

For each design implementation/description, indicate the following:

• Type of comparison being proposed (Differential (D) or Correlational (C)).
• The type of design (Not a design (N), Cross-sectional differential (XD), Cross-sectional correlational (XC), Case-control (CC), or Cross-sequential (XS)).

Design description/implementation	Type of comparison (D/C)	Type of design (N/XD/XC/CC/XS)
DESIGN #1. Recruit a single group of adults with a mental illness diagnosis and ask them to complete the CDOS and the MDS. Assess whether their responses to these two measures are related.	[   ]	[   ]
DESIGN #2. Use the MDS to identify and recruit a group of adults with a mental illness diagnosis who delay seeking medical assistance for a physical illness and compare their CDOS score to another group of adults (matched for age and type of mental illness) who do not delay seeking medical assistance for a physical illness.	[   ]	[   ]
DESIGN #3. Use the CDOS to form two groups of adults with a mental illness diagnosis – those who have experienced mental illness stigma in the medical context and those who have not. Compare the MDS scores of these two groups.	[   ]	[   ]
DESIGN #4. Use the CDOS to form two groups of adults with a mental illness diagnosis – those who have experienced mental illness stigma in the medical context and those who have not. Regularly measure and compare MDS scores of these two groups as they age.	[   ]	[   ]
DESIGN #5. Use the CDOS to measure mental illness stigma experienced in the medical context by one group of adults with a mental illness diagnosis and use the MDS to measure current personal delays in health seeking in a different group of adults with a mental illness diagnosis. Correlate CDOS scores of the first group with MDS scores of the second groups.	[   ]	[   ]

Hint: That’s right, one of these designs/implementations is a dud! For this design, it doesn’t matter whether you say it’s differential or correlational. Note also that a case-control uses a differential design (the DV is the grouping variable).

 

QUESTION 12

[All answers need to be correct to earn TWO marks for this question]

All of the four valid and relevant studies described in the list above are limited in terms of their internal validity (compared to a hypothetical experiment). But what advantage(s) do they have over any attempt at an experimental approach? (you can select more than one option)

[   ]             They have better external validity compared to an experiment

[   ]             They are more ethical because they do not attempt to manipulate person variables such as experiences of mental illness stigma

[   ]             They are more feasible because they do not attempt to manipulate person variables such as experiences of mental illness stigma

 

HYPOTHESES

You’re going to need hypotheses for these designs! We’ll focus on alternative hypotheses here. Remember, an alternative hypothesis needs to be relevant, concise and aligned with the design of the study! And it needs to predict a positive result. In other words, it is the result you’d obtain IF your predictions were confirmed. That’s the difference between the alternative hypothesis and the null hypothesis – the null hypothesis predicts that there will be no effect; the alternative hypothesis predicts that there will be an effect.

Select an alternative hypothesis for each design described in Question 1 (for the worksheet, perhaps use a line to connect a design with an appropriate hypothesis; for the weblink you’ll make the selection via a series of drop-down menus). Yes, there are many potential ways of wording these hypotheses – just pick the one that best matches each design. And yes, each design is uniquely matched to one and only one hypothesis.

QUESTION 13

[Each correct answer is worth 2/5^th of a mark. You can earn a total of TWO marks for this question]

DESIGN		ALTERNATIVE HYPOTHESIS
DESIGN #1. Recruit a single group of adults with a mental illness diagnosis and ask them to complete the CDOS and the MDS. Assess whether their responses to these two measures are related.		HYPOTHESIS #1. It is hypothesized that adults diagnosed with a mental illness who delay seeking medical assistance will be more likely to have experienced mental illness stigma in the medical context than similar adults who do not delay seeking medical assistance.
DESIGN #2. Use the MDS to identify and recruit a group of adults with a mental illness diagnosis who delay seeking medical assistance for a physical illness and compare their CDOS score to another group of adults (matched for age and type of mental illness) who do not delay seeking medical assistance for a physical illness.		HYPOTHESIS #2. It is hypothesized that adults who have experienced mental illness stigma in the medical context will report more delay in seeking medical assistance as they age compared to adults who have not experienced mental illness stigma in the medical context.
DESIGN #3. Use the CDOS to form two groups of adults with a mental illness diagnosis – those who have experienced mental illness stigma in the medical context and those who have not. Compare the MDS scores of these two groups.		HYPOTHESIS #3. It is hypothesized that adults with a diagnosed mental illness who have experienced mental illness stigma in the medical context will report more delay in seeking medical assistance than similar adults who have not experienced mental illness stigma in the medical context.
DESIGN #4. Use the CDOS to form two groups of adults with a mental illness diagnosis – those who have experienced mental illness stigma in the medical context and those who have not. Regularly measure and compare MDS scores of these two groups as they age.		HYPOTHESIS #4. It is hypothesized that mental illness stigma experiences will be positively correlated with personal delays in seeking medical assistance.
DESIGN #5. Use the CDOS to measure mental illness stigma experienced in the medical context by one group of adults with a mental illness diagnosis and use the MDS to measure personal delays in health seeking in a different group of adults with a mental illness diagnosis. Correlate the CDOS and MDS scores between the groups.		HYPOTHESIS #5. This is either not a logically valid design and/or not a conceptually valid design – there is no meaningful and/or testable hypothesis to match it.

 

MULTIVARIABLE RESEARCH

Simply observing an association between mental illness stigma and personal delays in seeking medical assistance doesn’t explain how and why these two variables are related. Psychologists are in the business of finding out!

 

Trust in medical professionals is known to promote medical help-seeking behaviour. Not surprisingly, the more trust in medical professionals, the more likely (and more quickly) it is that the patient will seek assistance from them. Therefore, you decide to include use the Trust in Doctors scale (TID) to measure trust and see what role it plays in the relationship between mental illness stigma and medical help-seeking delays. Let’s think through how this might work…

 

A colleague of yours believes that experiencing mental illness stigma in the medical context can reduce trust in medical practitioners, and it is reduction in trust that leads to subsequent delays in seeking medical assistance.

 

However, you believe that experiencing mental illness stigma in the medical context leads to delays in seeking medical assistance but only in people who already have low levels of trust in medical practitioners.

 

Yes, they sound like similar explanations, but they’re not!

QUESTION 14

Demonstrate that you understand how you and your colleague’s explanations differ by correctly identifying the different role that each variable plays in your studies. Please only select a single role for each variable listed (i.e., select only one option in each row), first in the context of your colleague’s study, and then in the context of your study.

[Each correct response is worth ½ a mark. You can earn a total of THREE marks for this question]

	IV/causal	DV/outcome	mediator	moderator
Variables in your colleague’s study
CDOS	[   ]	[   ]	[   ]	[   ]
MDS	[   ]	[   ]	[   ]	[   ]
TID	[   ]	[   ]	[   ]	[   ]
Variables in your study
CDOS	[   ]	[   ]	[   ]	[   ]
MDS	[   ]	[   ]	[   ]	[   ]
TID	[   ]	[   ]	[   ]	[   ]

QUESTION 15

[Each correct answer is worth 1/2 a mark. You can earn a total of THREE marks for this question]

Your task is to match each statement below with a term or concept from this unit. Each statement has a single best match. How to indicate which statement matches with which term? For the worksheet, you might want to draw a line connecting them. For the weblink, you’ll use dropdown menus.

 

STATEMENT		CONCEPT
STATEMENT #1. Mental illness stigma experienced in the medical context leads people to seek alternative therapies which in turn results in greater delays in seeking assistance from medical professionals.		#1. Procedural confound
STATEMENT #2. I discover a relationship between mental illness stigma experiences and subsequent delays in medical assistance seeking, but only for women, not for men.		#2. External validity
STATEMENT #3. I discover that when patients who have experienced DO answer questions from the CDOS, it can trigger feelings of distrust towards doctors (this does not happen to patients who have never experienced DO).		#3. Mediation
STATEMENT #4. I doubt that DO actually causes distrust of doctors. I think what might be happening is that people with low levels of pre-existing trust behave differently from others in at least two ways. First, they are more sensitive to mental illness stigma (whether or not it has actually occurred), and second, they are less likely to seek help from doctors when they feel unwell.		#4. Moderation
STATEMENT #5. I conduct a study into the effects of DO on perceived trust in medical professionals. Therefore, I compare trust levels in two groups – those with high CDOS scores and those with low CDOS scores. However, I’m worried that participants in my high-CDOS group also have more severe and more visible symptoms of mental illness (perhaps this is why they experienced DO in the first place!).		#5. Person confound
STATEMENT #6. My study into medical help-seeking in PWMI was criticised by a colleague because I only included people with mood disorders.		#6. Third variable

Note. Each statement is matched to one concept and one concept only and each concept receives only a single match (yes, that means you can use the process of elimination!).

 

 

 

YOU CAN NOW SUBMIT ASSESSMENT TASK 1

 

AT1 is worth 30%.

You will receive feedback and a mark (out of 30) for this assignment once the official submission date passes. If you wish to revise and resubmit you may do so via the same weblink you used for your original submission. Resubmissions are capped at 18/30 (60%) and students are not penalised if their resubmitted mark drops below their original mark.