Standardising experimental errorbars. What is appropriate?

38 Views Asked by At

I have a 5 replicates of an experiment comparing the amount of bacteria on various surfaces. I've standardise the values because the absolute values don't mean a lot to a potential reader: $\frac{x-\bar{X}}{S}$, where $\bar{X}$ is the sample mean and $S$ is the sample variance.

So far so good, but what should I do with the errorbars representing standard deviation units? How do these get standardised? The following figure shows the normalised values with error bars that are $\frac{s_i -\bar{S}}{S}$. Is this right? Does it make sense? Or is it non-sense? Data is below enter image description here

Location    Expt1   Expt2   Expt3   Expt4   Expt5
1   226 301 192 234 200
2   306 243 349 286 266
3   327 261 366 282 259
4   255 210 245 294 255
5   269 267 290 347 336
6   240 310 219 243 203
7   229 254 406 339 304
8   279 241 318 274 284
9   361 284 335 346 281
10  452 341 347 418 390
11  475 360 345 280 266
12  444 398 457 431 531
13  599 441 650 562 616
14  555 514 513 529 452
15  356 244 295 265 277
16  365 303 359 374 334
17  310 319 319 266 298
18  318 298 394 339 331
19  445 395 456 429 422
20  398 477 428 451 436
21  274 317 294 238 277
22  220 288 261 271 258
23  301 354 288 244 371
24  319 342 362 356 354
25  379 225 485 274 304