QAM Intro to Statistics, Data Collection, Sourcing.

LOCK IN 1

Statistics

• collection & describing data

• making interferences from samples 

• “Data Science”

Statistic

• single measure

• number used to summarize a sample data set

2 kinds of statistics

1. descriptive

2. inferential

descriptive statistics

collection, organization, presentation & summary of data (charts/ graphs)

inferential statistics 

uses a sample of data to draw conclusions and make generalizations about a larger population (estimating)

why study statistics?

to develop critical thinking, interpret research, make data-driven decisions in your career and daily life, and understand complex world issues like climate change and public health

statistical challenges

• imperfect data & practical constraints

• business ethics (upholding them) 

• using consultants (expensive = paid by hour, decisions faster if org knowns own statistics) 

• communicating with numbers (managers barely have time to read each numbers meaning & context)  

critical thinking 8 pitfalls

1. conclusion from small sample (not enough data)

2. conclusion from nonrandom samples (samples don’t represent population)

3. conclusion from rare events (ex: lottery eventually doesnt = win) 

4. poor survey methods

5. assuming a casual link (video game have violence = video game cause mass shooting) 

6. generalization to indiv (men taller than women) 

7. unconscious bias

8. significance vs importance (must be: significance = result is real, importance = result matters in a meaningful way) 

Data

collection of facts

diff between data & info

data = raw, unprocessed facts

information = data that has been processed

Data set

observation

• complete record of a single unit

• A single instance of the data being collected, essentially an individual case or entity. \

• each row in data set 

variables

• characteristic of subj

• each column in data set

• uni, bi, multi variate

data set

• all values we observe

• m x n or observation x variable 

Qualitative vs Quantitative data 

Qualitative = descriptions, experiences, and meanings

Quantitative = can be counted, measured, or expressed numerically.

4 types of data?

1. categorical

2. numerical

3. time series 

4. cross-sectional 

categorical data

• qualitative data

• for labeling

• nonnumerical values

coding - value of catgeorical variable represented using numbers 

binary - coding only has 2 values - arbitrary = choice is random

numerical data

• quantitative data

• for measurement

discrete - variable w/ countable no of distinct values (integers, whole numbers) “Can I count it one by one?”

continuous -  numerical value have any value with interval (any value within range, ex: 1.05) “Can I measure it more precisely?”

time-series data

• different equally spaced point in time 

• came from same unit; diff period in time 

• (Usually the bar/ line graph) 

cross-sectional data

• numerical value can have any value within interval

• came from diff unit at only 1 period in time

• any value within a range

• (scatter plot graph (circle thingy))

4 lvls of measurement 

1. nominal

2. ordinal 

3. interval 

4. ratio

Nominal measurement

• weakest

• codes used as placeholders no numerical meaning only for categories

• counting, mode 

• Ex: m=male , f=female 

ordinal measurement

• imply ranking of data value

• used to rank or order data into categories where the differences between the categories are not necessarily equal.

• Ex: educational attainment, income, etc

interval measurement

• rank has meaningful interval between scale points

• no meaningful zero 

• Ex: temperature, IQ score, SAT score, likert scale 

Ratio Measurement 

• strongest lvl of measurement 

• meaningful zero = represents absence 

• data can have negative no 

• Ex: weight, Height, Age, Time, kelvin tempt scale 

Precision

multiple attempts close to target

accuracy

hitting actual target

sampling concept: Population

all items we are interested in may be finite or infinite

sampling concept: sample

subset of population taken to analyze

selected members of the grp

sampling concept: census

examination of all items in defined population

every member of the grp 

situation where sample is preferred?

• infinite population

• accuracy

• timely results

• destructive testing

• cost

• sensitive info

situation where census is preferred?

• small population

• large sample size

• database exist 

• legal requirements 

sampling concepts 1: Parameter

• measure or characteristic of population

• μ = population mean

• π = population proportion

sampling concepts 2: Statistics

• numerical value calculated from sample

• x̄ = sample mean

• p = sample proportion

sampling concepts 1: Target population

• contains all indiv in which we are interested

• ex: population of those living in metro manila

sampling concepts 2: sampling frame

• grp from which we take the sample

• ex: names of ppl living in metro manila

Random sampling method 1: simple random sample

• every item in the population has same chance of being chosen 

  • sampling w/o replacement = once chosen remove from sample

  • sampling with replacement = can be called again 

Random sampling method 2: systematic sample

chosen every kth item from sequence starting randomly chosen entry among 1st k item

Random sampling method 3: stratified sample

within each stratum, simple random sample of desired size could be taken 

Random sampling method 4: cluster sample

taken from strata geographical regions

useful if:

• population frame & stratum characteristics not available rn 

• too expensive for stratified sample 

• some of loss of reliability is acceptable 

None random sampling method: Judgment

relies on expertise of sampler to choose items to represent population 

None random sampling method: convenience 

sample thats available 

None random sampling method: focus grp

panel of indiv chosen to represent wider population, form open-ended discussion & idea gathering 

Data Sourcing: primary source

raw info = gathered from 1st source in controlled/ uncontrolled situations 

Data Sourcing: secondary source

data acquired from optional sources = magazine, books, docs, etc.

• inner source = exist in stored orgs

• external source = gathered by other indiv from association’s other environment 

Survey types: mail survey 

• need targeted list

• expect low responses

Survey types: web 

• no bias 

• works best on targeted in well-defined interest grp on question of self-interest 

6 survey guidelines

1. planning

2. design

3. quality 

4. pilot test

5. buy-in

6. expertise

designing a questionnaire?

1. Open Ended Questions

2. Fill in the blank

3. Check boxes

4. Ranked choices

5. Pictograms

6. Likert scale

7. Short and concise instructions

8. Include an escape options (Others (pls specify))

9. Allow respondents to bypass sectors that are not relevant to them

designing a questionnaire what to look out for?

• multiple responses

• random replies for fill-in-the-blank

• range ans

• inconsistent replies