Efficient Data Collection

Efficient

Data

Collection

How can we be more efficient?

How can we be more efficient?

Why

Does

This

Matter?

1-10-100 Rule

The Cost of Quality: Data entry errors multiply costs exponentially based on stage at which identified & corrected

Prevention costs less than correction, which costs less than failure.

Where can errors be introduced?

Conduct a “Process Audit”

Process Audit 1: “Population Genetics of Lizards in Longleaf Pine Savannas”

After collecting lizards in traps, animals are returned to the lab and euthanized. A piece of each animal’s liver is removed, weighed using a scale, and stored in a plastic tube filled with ethanol. The data are later entered into a spreadsheet for analysis.

Process Audit 2: “Diversity & Dynamics of Tropical Tree Communities”

Every tree in a 10-ha plot is marked with an ID number and identified to species. The diameter at breast height (DBH) of each tree is measured with a tape measure and the location is recorded with a GPS. The DBH, species, and location of each tree are recorded on datasheets; in the evening at the field station these data are entered in a spreadsheet on a laptop computer.

Process Audit 3. “Economic Costs and Benefits of Diversified Crop Production in Rural Tanzania”

Villagers were interviewed about their agricultural practices. In addition to answering questions (asked by researcher with assistance from translator), participants were asked to make a ‘resource allocation map’: a drawing of the proportion of their income allocated to food, education, farming supplies, etc.

Audio recordings of the interviews were translated and transcribed to MSWord documents by bilingual students from the university. The maps were brought back to university so researchers could compare each subject’s interview responses with the budget allocations they drew on the map. Data about each interview (e.g., subject, location, translator, researcher name) were also entered in a spreadsheet).

What were the results of your process audit?

What can I do?

1. Checklists

2. Automation

3. Asset Management

4. UX/UI

1. Checklists

2. Automation

https://ifttt.com

3. Asset Management

Save time and reduce errors by labeling items (e.g., vials, sheets, forms) and “fill out” forms in advance.

4. UX/UI

Save time and reduce errors by formatting forms to speed up data collection, minimize errors, and streamline data entry.

1. Conduct a “Process Audit” of these data collection forms

2. What would your UX be if you were filling out these forms?

  • Reduce Cognitive Overload (aka keep it simple).
  • Write as little as possible

Use labels & keep them short…

Be careful with the location of input labels to avoid mistakes…

Poor alignment leads to filling out forms incorrectly

…and pay attention to where you put the labels

Use the appropriate input type and tag…

Radio Button vs. Check Box

…and format them in helpful ways

  • Blanks vs. Fill-in Boxes (narrow, wide, height)
  • Circling Options vs Check Boxes (distance between)
  • Likert scale options (vertical vs. horizontal, direction of impact, words vs. numbers)
  • Options to constrain errors vs. Unknown Results

Forms should be one column…

One-column forms are more comfortable to scan and conform better to mobile displays. Multiple columns can strain a user and cause them to skip fields accidentally.

…Except when multi-column makes sense

Fields that logically go together should be inline

Filled out faster and reduces cognitive overload

AVOID ALL CAPS

All caps make things hard and tiring to read

For online forms: Show all options if < 6 (unless space is limited)

1) Reduces Cognitive Overload (Dropdown selector requires 2 clicks)

2) Reduces Error (accidentally selecting incorrect option)

2) Reduces Error (accidentally selecting incorrect option)

Number questions and responses

Use Formatting to guide the data collectors

Efficient

Data

Collection