Comm Health Final Flashcards
The Null Hypothesis (Ho or sometimes NH)
null hypothesis refers to a general or default position: that there is no relationship between two measured phenomena, or that a potential medical treatment has no effect.
Correlation coefficient
a measure of the linear correlation between two variables X and Y
Value of correlation coefficient
between -1 and 1; closer to -1 or 1 is a stronger; 0 is no correlation
Positive correlation
graph with positive slope; as one variable increases so does the other or as one decreases so does the other - they both go the same way
Negative correlation
graph with negative slope; as one variable increases the other decreases or as one decreases the other increases - they go opposite directions
Strength of association
Value of r, can be -r or r; 0-0.25 is little if any association; 0.26-0.49 is weak; 0.7-0.89 is high; 0.9-1 is very high
p value
usually set at 0.05; probability that findings are due to chance; this is to test the hypothesis
Hypothesis testing
a statistical decision to reject or accept the null hypothesis based on probability (p) that is at a set significance level
Statistically significant
a value of p less than or equal to the set significant level means the results are statistically significant and you reject the null hypothesis
Not statistically significant
a value of p greater than the set significant level means the results are not statistically significant and you do not reject the null hypothesis
Type 1 statistic error
researcher rejects null hypothesis and concludes that a statistically significant difference exists when in fact no true difference is present
Type 2 statistic error
researcher concludes that no statistically significant difference exists and accepts the null hypothesis when in fact a significant difference does exist
Causes: type 1 error
large sample size; p value set too high (0.05 or greater); corrected by: randomly selected sample and good study design and set p value lower than 0.05
Causes: type 2 error
too small a sample size; unrealistic measuring devices; imprecise research methods; corrected by: correction of causes
t-test used in inferential statistics
statistical measure used when comparing hypothetical difference between TWO mean scores; NH for these tests - two unrelated groups are equal; looking to see if we can show that we can reject NH and accept alternative hypothesis - two NOT equal
ANOVA test used in inferential statistics
statistical measure used when comparing hypothetical difference between THREE OR MORE mean scores
Chi-Square test used in inferential statistics
*attitudes - non-numbers; used to analyze discrete, nominally scaled data, and to test differences between frequency distributions; test independence of two categorical (descriptive) variables; helps one arrive at a p value
Standard of acceptability
p < 0.05; 1 out of 20 occurred by chance or it has nothing to do with the testing situation
Validity
degree that a study or procedure measures what it claims to be measuring; is it measuring how it claims it should?
Sensitivity
the ability of a test to correctly identify the presence of a disease
Specificity
the ability of a test to identify the absence of a disease
Reliability
the extent to which the method of measurement consistently performs
Null hypothesis
a question or statement to be answered that can be stated in a negative outcome; no benefit or significance
research question or hypothesis
a question or statement to be answered; stated negatively or positively
Positive hypothesis
Brand X significantly whitens teeth
Negative (Null) hypothesis
No statistically significant difference exists between brand X and placebo; “NO SIGNIFICANT DIFFERENCE”
Dependent variable
the outcome of interest; change in it should be observed in response to some intervention
Independent variable
the intervention; what is being manipulated to change the outcome of interest
Extraneous variables
not related to the purpose of the study BUT may influence the outcome; interfere with accurate interpretation and produce invalid research results
Placebo
should have no positive or negative effects on subjects; should seem identical in every way to the real thing
Placebo effect
just knowing the desired effect produces it
Calibrated examiners
should be educated the same on how to perform; unification of what they do - do things the same
Intra-rater reliability
in one person; the one person is consistent in performance/ does the same all the time
Inter-rater reliability
between more than one person; they all have consistent performance/ do the same all the time
Hawthorne effect
they KNOW they are being watched and their performance is affected because of it - better
Examiner bias
they have a reason for which way the experiment goes; prefer a certain side and make sure their side comes out positively; examining their own or being paid by a biased company
Lack of control group
won’t have comparisons; less valid
Single blind study
subject does not know but examiner does know what is being tested; sometimes knowing examiners accidentally persuade results
Double blind study
both subject and examiner do not know what is being tested; *BEST outcome
Sample
portion of the population that, if properly selected, can provide meaningful information about the entire population
Simple random sample
gives every member of the population an equal chance of being selected; some mechanism of chance to choose them; no one is favored over another
Systematic sample
every nth individual participates; ex: if you have 1000 and want 100 of that, every 10th person
Stratified random sample
random sample carried further into sub groups; some from each group; if you want to be sure to cover such things as age, gender, income level, education levels, etc.
Biased sample
can lead to misleading results; sample chosen ensures results consistent with desired outcome
Judgement sample (Purposive)
choosing with a purpose for choosing certain groups; someone who knows the population chooses a sample to represent the population; risk of bias
Convenience sample
at convenience of researcher; simplest method; little concern for representativeness; may not be applicable to general population
Sampling error
occurs when a sample measurement is different from the population measurement; selecting sample not perfectly matched to represent entire population; can lead to inaccurate conclusions about the population
General sampling rules
30 subjects; study done for at least 6 months; should be of a meaningful population
Statistics
the science that deals with the collection, tabulation, and systematic classification of data
Types of data
Qualitative (categorical); Quantitative (continuous)
Qualitative
uses descriptive terms to measure or classify something; nominal, dichotomous, ordinal
Quantitative
uses *numerical values to describe something; interval, ratio
Nominal
variables with a name, that have no particular order; ie: puppy types, blood types, eye colors
Dichotomous
variables that only have TWO categories or levels; ie: gender: female or male, state of living: dead or alive
Ordinal
variables whose categories can be ordered or ranked, but the spacing between values may not be the same across the levels of variables; ordered but not proportionally ratioed; ie: education level (elementary, freshman, sophomore, junior, etc), pain level, cancer stages, *decay classifications (1-6)
Interval
points are equally spaced along the scale and the difference between the two points is meaningful (as opposed to ordinal scales); ie: temperature, IQ, pain scale 1-10, body length of infant; NO meaningful zero point
Ratio
ratios between points has meaning; Zero does count, can use “twice as much” rule; ie: age, weight, height, time, BP, distance, pulse, etc
Graphs
to express data; bar, histogram
Bar Graph
used to present categorical variables, bar for each category with spaces between to represent discrete nature of data
Histogram
similar to bar graph, but bars appear side by side (touching)
Measures of central tendency
single value to describe a set of data by identifying the central position within that set of data; mean, median, mode
Mean
arithmetic average of scores; most common measure of central tendency; particularly susceptible to extreme values; most useful and most familiar; always center of balance of distribution in a symmetrical distribution
Median
point of distribution with 50% of scores falling above it and 50% falling below it; NOT affected by extreme values; Midpoint: when total number is odd, median is midpoint; when total number is eve, take two middle scores and average
Mode
Most frequent score in a distribution; affects skew of graph, you CAN have 2 modes (bimodal), if everything is equal you have NO mode; least used of the measures of central tendency
Type of variable: Nominal
Best measure of central tendency: Mode
Type of variable: Ordinal
Best measure of central tendency: Median
Type of variable: Interval/Ratio (not skewed)
Best measure of central tendency: Mean
Type of variable: Interval/Ratio (skewed)
Best measure of central tendency: Median
Measures of dispersion (spread)
Range; Variance; Standard of Deviation; used to describe how much variation is present in a sample
Range
difference between the highest and lowest scores in a data set; simplest measure of a spread; Range = Maximum value - Minimum value; ie: 95K-12K=83K; smaller (narrow) range is better because a large range means the mean is not as representative of the data
Variance
represents the average distance of each score from the mean
Standard deviation
the square root of the variance; a measure of the spread of scores within a set of data; appropriate when data isn’t skewed or has outliers;
Bell curve
most of the time standard deviation works within a normal bell curve distribution
Positive (RIGHT) skew
outliers create positive skew when most scores are lower but one or two are higher ——–>
Negative (LEFT) skew
outliers create negative skew when most scores are higher but one or two are lower <——-
Skewed distribution
distribution of scores is NOT symmetrical
Normal distribution
bell curve; 68-95-99.7 rule; the majority of the scores always falls within +1 or -1. That is a given. It will always be true regardless of the value of the standard deviation