Sign in to follow this  
Followers 0
drphoenix

Stats and research methods

14 posts in this topic

Which of the following is a discrete variable

Choose one answer. a. Annual income

b. Blood pressure

c. Age in years

d. Years of education

e. Gender

Answer E

[highlight]A discrete, otherwise known as categorical, variable can only have set values (of whole numbers[/highlight]) and can be subdivided into : binary, nominal, ordinal, interval or ratio data.

What????? Birmingham says the above

Superegocafe notes says Categorical is nominal and ordinal

And that Discrete and Continuous are quantitative i.e. according to superego Discrete is not categorical.

Whats going on?

Can anyone see what i am talking about?

Share this post


Link to post
Share on other sites

Discrete variables are counts, they have a value.

Categorical variables do not have a value, although there may be an inherent order of the categories.

This question doesn't make sense.

Share this post


Link to post
Share on other sites

Thank you webmaster.

That question really rocked my boat. After all the reading, its just unforgivable for any course to make errors like that, especially for those who only really on course notes like me.

Share this post


Link to post
Share on other sites

Dear webmaster,

Could you do us a huge favour and sort this out for me please. Can you give me all the possible examples for the different variables, in the context of the exam please? I've realised that if it stick to my formula, I could solve most questions, however I realise that knowing which variables are being talked about is crucial to use my formula. So, here goes,

1.RATIO

This would be height and Weight

2.INTERVAL

This would be Temperature and B.P

3.CONTINUOUS

This would be Age, ?Height and ?Weight

4.ORDINAL

Anything that is rated. i.e. Likerts; ranking;Rating etc

5.NOMINAL

Gender; Marital status; Diagnosis

6.BINARY

No idea.

I am using fast and dirty rules here to remember, hence the simplicity.

Thanks in advance. I know I have a long way to go but I see this section as valuable enough to make the difference between me passing or failing.

Share this post


Link to post
Share on other sites

Could you give me a PO dose of your dirty quick rules of statistical method

Share this post


Link to post
Share on other sites

I will in a bit Eamon [not sure ifthey will work, though]. But first here are a few more questions that i need help with

A researcher gives out 100 questionnaires and ask subjects ten questions about how sad certain things make them. They are provided with a separate test to measure their stress level. The researcher is interested to find mathematically the best equation to predict stress levels from the ten behaviours. The statistical test they should be using would be

a. ANNOVA

b. Multiple linear regression

c. Linear Regression

d. t test

e. Logistic regression

Share this post


Link to post
Share on other sites

In a RCT, a group of patients were given a new drug for anxiety. They were asked to rate whether they felt better or not at 2 weeks and subsequently at 4 weeks. The test to compare the difference in how they felt at 4 weeks and 2 weeks is

a. Mann-Whitney U test

b. McNemar test

c. Kruskal Wallis test

d. Friedman test

e. Chi-squared test

Share this post


Link to post
Share on other sites

A.     If Parametric data is given i.e Ratios / Intervals / Continous and

>2 groups    = ANOVA

 2 groups     = Independant t test

    Paired t test

B.   If ?Non-Parametric data is given and

>2 Groups   = Kruskal wallis

 2 Groups    = Mann Whitney

   Willcoxon

C.   If Non-parametric i.e. categorical is given and

> 2 groups   = Chi square

  2 groups    = Chi square

    McNemar (Binary data)

Thats what i have understood so far and it worked 8 out of ten times on Birmingham.

Share this post


Link to post
Share on other sites
In a RCT, a group of patients were given a new drug for anxiety. They were asked to rate whether they felt better or not at 2 weeks and subsequently at 4 weeks. The test to compare the difference in how they felt at 4 weeks and 2 weeks is

 a. Mann-Whitney U test    

 b. McNemar test    

 c. Kruskal Wallis test    

 d. Friedman test    

 e. Chi-squared test  

[highlight]

McNEMAR TEST[/highlight]

Share this post


Link to post
Share on other sites
In a RCT, a group of patients were given a new drug for anxiety. They were asked to rate whether they felt better or not at 2 weeks and subsequently at 4 weeks. The test to compare the difference in how they felt at 4 weeks and 2 weeks is

 a. Mann-Whitney U test    

 b. McNemar test    

 c. Kruskal Wallis test    

 d. Friedman test    

 e. Chi-squared test  

[highlight]

McNEMAR TEST[/highlight]

Hey idle, I was expecting an explanation man. I know thats the answer from Birmingham but why not Chi?

Share this post


Link to post
Share on other sites
In a RCT, a group of patients were given a new drug for anxiety. They were asked to rate whether they felt better or not at 2 weeks and subsequently at 4 weeks. The test to compare the difference in how they felt at 4 weeks and 2 weeks is

 a. Mann-Whitney U test    

 b. McNemar test    

 c. Kruskal Wallis test    

 d. Friedman test    

 e. Chi-squared test  

[highlight]

McNEMAR TEST[/highlight]

Hey idle, I was expecting an explanation man. I know thats the answer from Birmingham but why not Chi?

IT'S CATEGORICAL DATA..............FEELING BETTER OR NOT...........THAT'S IT...........SO, I WOULD NOT GO FOR CHI-SQUARED TEST AS McNEMAR'S IS FOR PAIRS.......( AT 2 WEEKS & 4 WEEKS)............

MY TIP...............BLINDLY BY HEART THE TABLE FROM B'HAM COURSE......

that helped 2 answer all da Qs

Share this post


Link to post
Share on other sites

Although we assume the outcomes to be dichotomous, mostly we measure improvement on scales or even the CGI is kind of a likert scale...

Your rationale for Mcnemar test is based on he fact it is a 2X2 contingency table. So the table would be

Improvement at 2weeks vs Improvement at 4 weeks

4 weeks

+ -

2 wks +

-

Which is fine....

But if you were measuring the anxiety ratings on the scale, and comparing it with baseline, 2 weeks and 4 weeks, the test would be a Friedman test...

Btw... I think the answer would be McNemar in this question...

Share this post


Link to post
Share on other sites
Which of the following is a discrete variable

Choose one answer.   a. Annual income    

 b. Blood pressure    

 c. Age in years    

 d. Years of education    

 e. Gender    

Answer   E

[highlight]A discrete, otherwise known as categorical, variable can only have set values (of whole numbers[/highlight]) and can be subdivided into : binary, nominal, ordinal, interval or ratio data.

What????? Birmingham says the above

Superegocafe notes says  Categorical is nominal and ordinal

And that Discrete and Continuous are quantitative i.e. according to superego Discrete is not categorical.

Whats going on?

Can anyone see what i am talking about?

A discrete variable is one that cannot take on all values within the limits of the variable. For example, responses to a five-point rating scale can only take on the values 1, 2, 3, 4, and 5. The variable cannot have the value 1.7. A variable such as a person's height can take on any value. Variables that can take on any value and therefore are not discrete are called continuous.

Statistics computed from discrete variables have many more possible values than the discrete variables themselves. The mean on a five-point scale could be 3.117 even though 3.117 is not possible for an individual score.

A discrete variable, as contrasted with a continuous variable, has distinct categories. A discrete variable with only two categories is a dichotomy. If it has more than two, it is a polychotomy.

Discrete variables also are of two types: unorderable (also called nominal variables) and orderable (also called ordinal).

So in the question, there are two answers... that cannot take any decimel values...

For eg: Gender can take only two values- it can only be male (1) or female (2), you cannot have a 1.5 as an individual response. All qualitative variables are discrete.

BP cannot take any value - like decimel values for eg: BP cannot 120.5/ 80.2 (so that is discrete technically... but if there was a machine which gave us BP measures in decimels, then it is not)

All other variables are continuous - age, years of education and annual income can be anything in the range...

So.... the best answer would be Gender... I would say

Does that make sense????

Share this post


Link to post
Share on other sites

Candidates should note that for the June 2009 paper onwards the Paper 2

‘Statistics and research (basic)’ section will move from the Paper 2 to the Paper 3.

Announcement on the College website.

Share this post


Link to post
Share on other sites
Sign in to follow this  
Followers 0