selectby tag

To use selectby: condition, options should include a condition decorator. In the example below, each respondent is assigned a color category in QCOLORTYPES based on their selection in Q4.

4. Which color is your favorite?
type: radio
1. black
2. gray
3. blue
4. green
5. purple
6. red
7. pink
8. orange
9. yellow

COLORTYPES. Color categories
type: radio
selectby: condition
1. Neutral colors { if anyChecked($Q4,1,2) }
2. Cool colors    { if anyChecked($Q4,3..5) }
3. Warm colors    { if anyChecked($Q4,6..9) }

To use selectby: random, options can simply be listed on the variable. If no conditions are applied, the software chooses an option at random. 

The example below selects three options at random. However, it also uses the exclusive tag to tell the selectby logic which options can't be picked at the same time.  

• If 1 is selected, then can't select 2 (and vice versa)
• If 3, then can't select 4 (and vice versa)
• Also if 3, can't select 5 or 6 (but 5 can have 6 and vice versa)

TRACK. Pick three
type: checkbox
selectby: random
maxgroups: 3
exclusive: 1,2
exclusive: 3,4
exclusive: 3,5,6
1. Apples 
2. Oranges 
3. Pear
4. Carrot
5. Tomato
6. Watermelon
7. Cherries
8. Canteloupe
9. Jackfruit

The example below uses selectby: counts at QCOLOR. First, an option must be selected at Q1 to be eligible. Then, the system assigns the respondent to the option with the lowest number of completes at the time of response.

set list: COLORS
1. Red
2. Orange
3. Yellow
4. Green
5. Blue
6. Indigo
7. Violet

1. Which of the following colors do you like? 
instructions: Check all that apply.
type: checkbox
optsfrom: COLORS

COLOR. Single color selection
type: radio
selectby: counts
optsfrom: COLORS {if anyChecked($Q1,[id])}

To select option(s) based on a priority or weighted value defined in the survey code, you should use selectby: weight. The highest weight is selected first. If no weight is defined, the default value is '0.' Options with a weight of '0' can still be selected unless excluded. As with all methods, options can only be selected if the conditions are met, regardless of the weight.

Examples:

• Static weights: Always prioritize options based on a given importance.
• Response-based weights: Use respondent answers, such as numeric inputs, as weight values.
• Conditioned weights: Logical conditions (e.g., anyChecked) are evaluated as '1' if true and '0' if false. They can act like values in a weight decorator to adjust weights dynamically based on previous responses.

Static weights

In this example, if option 1 (McDonald's) was selected at Q1, a weight of '10' is assigned. If option 6 (Subway) was also selected, a weight of '1' is assigned. The respondent would be assigned McDonald's at QFOOD because it has the highest weight between those two eligible options.

set list: FASTFOOD
1. McDonald's
2. Wendy's
3. Burger King
4. In-N-Out
5. KFC
6. Subway

1. At which of the following fast food restaurants have you eaten in the past six months?
type: checkbox
optsfrom: FASTFOOD

FOOD. Food quota, prioritizing McDonald's and Burger King, then Wendy's and In-N-Out, and finally KFC and Subway
type: quotas
selectby: weight
optsfrom: FASTFOOD [1,3] { if anyChecked($Q1,[id]) } {weight: 10}
optsfrom: FASTFOOD [2,5] { if anyChecked($Q1,[id]) } {weight: 5}
optsfrom: FASTFOOD [4,6] { if anyChecked($Q1,[id]) } {weight: 1}

Response-based weights

In the example below, the numeric response at Q1 populates the placeholder '$Q1R[id]' for the same option at the QFOOD quota. For instance, if Subway is rated '8' and KFC is rated '2' at Q1, each option will have that value inserted as the weight at QFOOD, with all other options set to '0.' The QFOOD quota would assign the respondent to Subway because it has the highest weight. In the case of a tie, the fast food restaurant with the fewest completes (counts) will be assigned.

set list: FASTFOOD
1. McDonald's
2. Wendy's
3. Burger King
4. In-N-Out
5. KFC
6. Subway

1. Of the last ten times you have eaten fast food, how many have been at the following restaurants?
type: integer table
range: 0-10
rowsfrom: FASTFOOD
sct: Total {total: 10}

FOOD. Food quota, prioritizing based on where the respondent has eaten most in the past year
type: quotas
selectby: weight, counts
optsfrom: FASTFOOD {weight: $Q1R[id]}

Conditioned weights

Using the FASTFOOD list again, the condition if anyChecked($Q1,[id]) ensures that only options "ever heard of" at Q1 are eligible for selection at QFOOD. Since Q2 and Q3 depend on Q1, no additional conditions are needed.

Weights prioritize selection at Q3 as the most important, followed by Q2. There is no need to assign a weight for Q1, as it holds the lowest priority.

1. Which of the following fast food restaurants have you ever heard of?
type: checkbox
optsfrom: FASTFOOD

2. And at which have you ever eaten?
type: checkbox
optsfrom: Q1

3. And at which have you eaten in the past 12 months?
type: checkbox
optsfrom: Q2

FOOD. Prioritizing based on any restaurants frequented in the past year, followed by any ever eaten, then if ever heard of
type: quotas
selectby: weight, counts
optsfrom: FASTFOOD { if anyChecked($Q1,[id]) } {weight: anyChecked($Q3,[id])*100 + anyChecked($Q2,[id])*10}

In some cases where selectby: weight is used, selectby: order can be used instead to achieve the same result. The system checks each option in the order they are listed until it finds one that meets the condition. 

In this example, if 'anyChecked($QCURRENT,1)' is true for the respondent, they are assigned option 1 "Purchaser." If not, the system moves down to option 2. If 'anyChecked($QFAMILIAR,1)' is true for the respondent, they are assigned option 2 "Aware." If neither option 1 nor option 2 is true for the respondent, the system defaults to option 3, which has no condition and is available to all respondents.

USER_TYPE. User type overall
type: radio
selectby: order
1. Purchaser { if anyChecked($QCURRENT,1) }
2. Aware     { if anyChecked($QFAMILIAR,1) }
3. Not aware

To use selectby: calculate, apply the cvalue tag to perform the necessary calculations used to select options.

In the example below, the QSMP variable utilizes the url_param method to pull in information from the URL parameter 'smp' in a survey entry link. If the value '10', '20', or '30' is found, it will be selected at QSMP; otherwise, '0' will be the default selection.

SMP. SAMPLE via URL-parameter
type: radio
selectby: calculate
cvalue: url_param('smp') || 0
0. NO "smp" parameter-value
10. Representative
20. Augment
30. Oversample

To use selectby: percentage, quota targets must be set for all applicable quota buckets in the Quotas applet.

For example, if researchers aim for 50 responses each for BBC and CMT, and 100 each for Comedy Central and Discovery Channel, these maximums should be entered on QDEEPDIVE in the Quotas applet. As respondents answer Q3, they will be assigned to the channels with the lowest percentage of progress toward the specified maximums. This ensures that all buckets will fill evenly, despite having different maximums.

3. Which of the following TV networks do you watch at least once a week for an hour or more?
type: checkbox
1. BBC America
2. CMT
3. Comedy Central
4. Discovery Channel
5. FOX
6. ESPN

DEEPDIVE. QUOTA | 2 Brands for Deep Dive
type: quotas
maxgroups: 2
selectby: percentage
optsfrom: Q3

The weighted_draw method allows you to specify a proportional filling scheme among options.

Imagine adding candies into a jar. You add different amounts of each type of candy as defined by its weight. Then, you randomly pick candies from the jar.

This method allows options to be assigned based on known percentages of total respondents, without necessarily knowing how many total respondents there will be. 

It's important to attach the weight decorator to each option, as items without an assigned weight will never be selected.

PQ. Product quotas
type: radio
selectby: weighted_draw
1. Product A {weight: 4}
2. Product B {weight: 2}
3. Product C {weight: 1}

In the above example, Product A is twice as likely be selected as Product B, which is then in turn twice as likely to be selected as Product C. The total weighted_draw values sum to 7. So, each product's chance can be thought of as a fraction: Product A has a 4 out of 7 chance, Product B has a 2 out of 7 chance, and Product C has a 1 out of 7 chance of being chosen.