MDM4U- Qualitative Data, survey methods, bias

qualitative data

data representing concepts that cannot be represented by numbers

frequency table

Shows the number of individuals for each data point

how do you calculate a proportion?

divide each count by the number of data values.

multiply by 100 for a percentage

why should you represent data in a graph?

• relationships/patterns can be seen clearly

• show important features (greatest, smallest)

• allow for greater comparisons

• patterns can be revealed

simple random sampling

• researcher randomly selects a subset of participants from a population

• everyone has equal chance of being selected

systematic random sampling

• samples chosen based on a system of intervals

• researchers set a sample size and interval number

• start at a random number, continue with the chosen interval until the sample size has been fulfilled

stratified random sampling

• used to accurately represent all subgroups (strata)

• strata can be separated from the population by age, gender, ethnicity, etc)

• equal percentage of people from each strata

cluster random sampling

• divide population into random groups (clusters) based on factors (age, gender, etc)

• pick a random cluster, collect data from everyone in it

multi-stage random sampling

• divide population into smaller hierarchical groups (principal, teacher, students)

• A predetermined number of people are chosen from each group

convenience sampling

• only sample individuals who are convenient to the researcher

• focuses on location and avalibility of sample

voluntary response sampling

• only take info from people who agree to participate

sample bias

• when data set is not representative of the entire population

household bias

• when one group is overrepresented due to the size of the population

• (there are more g9s, so they will be a large portion of results)

measurement bias

when method of collecting data is under or overestimating a characteristic

leading question bias

• lead the respondent to a certand answer

• “how helpful was our fantastic customer service?”

loaded question bias

• assume somehting about respondent

• “do you agree ______”

double-barrelled question bias

• pose questions about more than one topic

• “how much do you enjoy biking and swimming”

observer/researcher bias

• when method of observation results in differences from reality

testing bias

• data is not equally accurate to each group

respondent bias

when the responses are dishonest, particularity bc of peer pressure

recall bias

• when participants do not remember past events that could be important to the survey

non-response bias

• person’s refusal to answer the question

response bias

• when ppl do not answer the question honestly

acquiescence bias

tendency to say yes to every question

demand characteristics bias

cues that might indicate the research objectives to participants

social desirability bias

tendency to change answers so they are in line with societal expectations

courtesy bias

tendency to be polite or corteous towards the researcher

question-order bias

reacting differently to questions based on how they were ordered

extreme responding bias

tendency to only choose highest or lowest response

manipulating the y axis

• when the graph has a y axis that does not start at 0

• create the impression of a large change when there is none

area principle

• when amounts are compared by creating images, the area of the images must be proportional to the amounts

• if one amount is twice as much as another, it must be twice as large

inverted axis

• y axis has been flipped to create the ‘opposite’ illustration

proper unit of measurement

• some measurements should be represented using ratios (per 100 ppl), especially when comparing populations