What is variance?

Why are the difference squared?

It is the difference b/w the individual values in a set of data and the mean 

Because their are positive and negative variables.

T or F

The descriptive measures of a sample are its parameters. The descriptive measures of a population are statistics.

False

The descriptive measures of a sample are its statistics. The descriptive measure of a population are its paramters.

What is a sample?

What do they use statistics for?

Give examples of descriptive measures for statistics

A subset of a larger population that

is of manageable size.

Samples use

statistics to determine larger populations. 

Mean (x-bar), standard deviation (s), variance (s²)

How is a Confidence level different than a CI?

A CL shows how likely an interval (a CI) will include a parameter. A CI is an interval that is likely to include a parameter. 

T or F

As the confidence level desired decreases,

the CI widens.

False

As the CL desired increases, the CI widens.

What is sd?

It is a measure of the spread of values.

It is measured by calculating the square root of variance.

Give examples of descriptive measures for parameters.

mean (m), standard deviation (s) & variance (s²).

Is an exam letter grade C or D?

Is it N, O, I, or R?

It's discrete b/c it's ordinal. It's ordinal, b/c it is a ranking and the difference b/w exam grades is not = (you could have a 75 for a C, a 81 for a B, & a 97 for an A)

T or F:

The differences b/w levels need not be = to be classified as ordinal.

True

T or F:

Temperature on a fahrenheit scale is considered Interval.

True, b/c it can include non-integer values (55.6⁰F), the differences b/w its levels are =, and it doesn't have an absolute zero (molecules are always moving).

Which category does pain scale (1-10) fall under?

Is height C or D? Is it N, O, I, or R?

Is bp C or D? Is it N, O, I, or R?

Is occurrence of stroke C or D? Is it N, O, I, or R?

Is Lives saved C or D? Is it N, O, I, or R?

C and R

C and R

D and N (not sure about N)

D and O (not sure about O)

What is the mode?

What is the mean?

What is the median?

The mode is the most frequently occuring number. You can have more than one mode.

The mean is an average.

The median is the middle value when you line up data in rank order. If the middle value is 2 even #'s, just average them.

Measures of central tendency, charts, and graphs are examples of                  

What are the measures of central tendency?

Descriptive measures

Mean, median and mode

Notice that only I and R use mean. That's why you ask yourself, "can i take a meaningful mean of this data" to determine the level of measurement. 

N - mode only

O - mode and median

I - mode, median, and mean

R - mode, median, and mean

What do measures of variability do? 

What are the three measures of variability?

They describe the degree to which observations vary.

Range, variance, and standard deviation.

The range is the difference b/w the highest and lowest values, but it's rarely used. Variance is the difference b/w individual #'s and the mean. sd is a measure of the spread of values.

T or F:

If data values are close together, the differences b/w the observations (x) and the mean of the observations (x-bar) are small, which implies a small variance and a small variability.

T or F:

A small standard deviation means that the data points are all close to the mean, and a large standard deviation means that all of the data points are far from the mean.

What do standard deviations do?

What % of values will lie w/in one standard deviation (s) of the mean? W/in 2 s? W/in 3s?

Essentially, they tell us how far data points are from the mean.

68%; 95%; 99.7%

Calculate s and s² for the following sets of data:

Data: 50, 54, 46, 45, 56, 48, 52

Data: 28, 34, 22, 26, 30

For the first set: s² = 16.80 & s = 4.1

For the second set: s² = 20 & s = 4.47

T or F:

Variability w/in subjects tends to be greater than variability b/w subjects.

False

Variability b/w subjects tends to be greater than variability w/in subjects.

T or F:

A distribution that is skewed will not affect the mode, will slightly affect the median, and will greatly affect the mean.

A distribution that is skewed to the right is                  affected. A distribution that is skewed to the left is                affected.

True

positive; negative

T or F:

In a normal distribution curve, the tail never touches the x-axis.

T or F:

In a standard normal distribution curve, the

median = mean = mode

True

True

zero; one

T or F:

A small sd will have a small, flat peak w/ longer tails.

False 

A large sd will have a small, flat peak w/ longer tails. A small sd will have a sharp, increased slope w/ shorter tails.

T or F:

Since most distributions are not normal, any normal distribution can be standardized to a normal distribution by calculating a t-value.

False

It's a z-value, not a t-value.

Standard normal deviations are                 and                  .

T or F:

Statistical normality has nothing to do w/ medical normality

symmetrical unimodal

True

Standard normal distributions and z-scores can only be used for                 &                    .

1) Each x is 0 or 1 (ie, 0 = live & 1 = die; or

0 = pass & 1 = fail)

2) Each x is independent of the other (ie, just b/c one person dies doesn't mean the other will)

3) The probability that x = 1 is the same for every x

T or F: 

Any time a result is statistically significant, it will also be clinically significant.

T or F:

You shouldn't ever say something is approaching significance.

False

You have to use your professional judgement to decide.

True

It may be avoiding significance. You don't know.

What is a random variable?

T or F:

Independent or dependent variables can be random.

It's something that can take any potential quantity.

True

However, we hope that dependent variables are not random. Instead, we hope that they are dependent on the independent variable.

True or False:

Random variables are discrete variables if there are no gaps in between. Examples include blood pressure and blood sugar.

False

Random variables r discrete variables if there r gaps b/w values. Fe, if a male is assigned #1 & a female #2, there would b no 1.2 or 1.3, etc. A continuous variable, otoh, would include all the values between 2 values. Examples include bp & blood sugar.

T or F:

The peak is what we use to determine probabilities.

False

The distribution of a curve is what we use to determine probabilities.

T or F:

Single measures on the different patients are independent, but multiple measures on the same patient are dependent.

T or F:

If a z-score ends up in the tail of distribution, the outcome didn't happen by chance.

T or F:

If you had 2 distributions with the same mean, but different sd's, the two curves would have the same shape, but their peaks would be at different locations.

False

If you had 2 distributions with the same sd and different means, the two curves would have the same shape, but their peaks would be at different locations.

T or F:

If the sd of a distribution is > than the mean, the distribution is non-standard and there is a lot of dispersion.

True

I'm confused on this one, b/c the sd is 1 and the mean is 0 for a standard normal distribution. In this standard normal distribution, the sd is > than 1, so how is this non-standard.

How do you know whether to use sd or SEM?

T or F:

z-scores are used for smaller samples

False

t-tests are used for smaller samples.

What are t-scores?

What are 2 things t-scores are used for?

t-scores are a distribution of sample means

1) to compare a small, sample mean to a large, known population

2)to compare two samples to each other and make inferences.

T or F:

A larger # of people will equate to a smaller SEM and a sharper peak, b/c the distribution is less spread out.

True

However, this doesn't make sense to me, mathematically. The SEM is in the denominator, so shouldn't the t-score get bigger and have a flatter peak when the SEM gets smaller?

It is the # of observations that can be chosen freely once the mean has been established. If a

mean is fixed, it is not allowed to be free. So, if you have one mean, only n-1 can be free (where n is the # of observations); you lose 1df. If you have 2 means, you lose 2 df's.

x-bar = 79, mu = 83, SEM = 1.02, n = 124

What is t?

Is this sample mean w/in 95% of all calculated sample means for this population mean?

t = -3.91

No, b/c 95% only contains 1.9799

Why do we have to use CI's? Why can't we just use t-scores?

It is the amount of confidence that the CI contains the population parameter. CL's are all about replications. If a CL is 95%, this means that you can replicate a certain result 95 times out of 100.

If a CI is 95%, it means that we are 95% sure that the interval contains the population parameter.

T or F:

Z-scores r sufficient 4 finding population distribution, b/c populations r small enough that the SEM stays constant.

False

Z-scores r sufficient 4 finding population distribution, b/c populations r large enough that the SEM stays constant.

t-distributions have                peaks

and               tails than s-distributions.

smaller; fatter

T or F:

Increasing df would increase the resemblance of a t-distribution to standard normal distribution.

T or F:

As the CL decreases, the CI widens & the t-value increases

T or F:

As the t-value increases, it increases w/in the rows (from left to right)

False

As the CL increases, the CI widens.

True

1) Random sampling from a robust sample

2) Samples (drug A & drug B) must be independent

3)There must be normal population (meaning you need a large sample. If your sample is too small, your data will be skewed)

Pooled variance, paired samples, and separate variance are all methods used to find                             ?

What are 4 "very important assumptions" for a pooled variance CI?

CI for a difference between means

1) Random sampling

2) Must have independent samples(drugA & drug B)

3) Must have independent observations

4) Must have homogeneity of variance (meaning both samples must have = variance. If they don't, you have heterogeneity of variance and you have to use the separate variance method).

T or F:

If 1 is included in the CI, there is no difference b/w means.

False

If 0 is included in the CI, there is no difference b/w means.

Quinapril

HTN

20 - 80 mg qd